mass spectrometry-based drug accumulation assay in Pseudomonas aeruginosa

Analytical Biochemistry 385 (2009) 321–325

Contents lists available at ScienceDirect

Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio

Development of a liquid chromatography/mass spectrometry-based drug accumulation assay in Pseudomonas aeruginosa Hongliang Cai a,*, Kelly Rose a, Lan-Hsin Liang b, Steve Dunham c, Charles Stover c a

Department of Pharmacokinetics, Dynamics, and Metabolism, Pfizer Global Research and Development, Michigan Laboratories, Ann Arbor, MI 48105, USA Molecular Technology, Pfizer Global Research and Development, Ann Arbor, MI 48105, USA c Antibacterial Pharmacology, Pfizer Global Research and Development, Ann Arbor, MI 48105, USA b

a r t i c l e

i n f o

Article history: Received 26 September 2008 Available online 5 November 2008 Keywords: Bacterial penetration Bacterial accumulation Bacterial efflux Pseudomonas aeruginosa LC/MS Ciprofloxacin

a b s t r a c t Bacterial resistance to antibiotic therapy remains a worldwide problem. In Pseudomonas aeruginosa, rates of efflux confer inherent resistance to many antimicrobial agents, including fluoroquinolones, due to a high level of expression and a relatively high turnover number of the efflux pumps in gram-negative bacteria. To understand the roles of efflux pumps in both the influx and efflux of compounds in P. aeruginosa and to aid the chemistry compound design by bridging in vitro enzymatic binding data (IC50 values) with whole cell results (MIC numbers), a collaborative effort was put forward to validate a series of bacterial penetration/accumulation assays for assessment of intracellular drug concentration. Initially, using 2-(4dimethylaminostyryl)-1-ethylpyridinium cation (DMP) as the tracer, a 96-well fluorescence assay was established to measure the time-dependent accumulation of DMP in wild-type (PAO1), MexABOprM deletion (PAO200), and MexABOprM–MexCDOprJ–MexJKL:FRT deletion mutants (PAO314). At steady state, the order of DMP accumulation was PAO314 > PAO200 > PAO1. Subsequently, the established assay conditions were applied to a radiolabeled assay format using 3H-labeled ciprofloxacin. At the concentration tested, the accumulation of [3H]ciprofloxacin approached a plateau after 15 min and the amount of accumulation in PAO314 was higher (2- to 10-fold) than that in PAO1. Finally, with an additional step of cell lysis, a liquid chromatography/mass spectrometry-based assay was established with ciprofloxacin with (i) superior sensitivity (the detection limit can be as low as 0.24 ng/ml for ciprofloxacin) and (ii) the ability to monitor cold or nonfluorescent compounds in a drug discovery setting. Ó 2008 Elsevier Inc. All rights reserved.

For antibiotic drug discovery efforts focusing on novel targets, making the step from in vitro target enzyme inhibition to whole cell activity is very difficult. A significant challenge in conquering this hurdle lies in the fact that the structure–activity relationships (SARs)1 for bacterial penetration and efflux avoidance are poorly understood [1,2]. The underlying reasons can be twofold: (i) compounds can be actively excreted from cells via efflux pumps and (ii) the physicochemical properties of compounds do not favor bacterial penetration. Therefore, it would be beneficial to independently measure an experimental compound’s ability to penetrate and persist in the target bacteria irrespective of whole cell inhibitory activity. This is particularly true in Pseudomonas aeruginosa, a type of gram-negative bacteria possessing several efficient efflux systems

* Corresponding author. Fax: +1 314 274 4426. E-mail address: hongliang.cai@pfizer.com (H. Cai). 1 Abbreviations used: SAR, structure–activity relationship; LC/MS, liquid chromatography/mass spectrometry; DMP, 2-(4-dimethylaminostyryl)-1-ethylpyridinium cation; CCCP, carbonylcyanide m-chlorophenylhydrazone; OD, optical density; FLIPR, Fluorometric Imaging Plate Reader; MS/MS, tandem mass spectrometry; ESI, electrospray ionization; MRM, multiple reaction monitoring; MIC, minimal inhibitory concentration; PK, pharmacokinetic; PD, pharmacodynamic. 0003-2697/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2008.10.041

that can export a wide variety of molecules, including antibacterial drugs, imparting a generally high level of intrinsic antibiotic resistance [3,4]. To this end, multiple techniques have been employed to assess intracellular drug concentrations [5–12]. All of the methodologies were based on either fluorescent or radiometric detection, neither of which is amenable to the drug discovery setting due to the limited availability of radiolabeled compounds and lack of fluorescent moiety for many novel antibacterial discovery compounds. To meet the drug discovery needs, an alternative approach would be to set up a liquid chromatography/mass spectrometry (LC/MS)-based bacterial penetration assay in both the efflux pump knockout P. aeruginosa PAO314 to track intrinsic bacterial permeation and an isogenic wild-type strain (PAO1) to assess efflux and penetration. In this article, we present an LC/MS-based methodology to measure cellular ciprofloxacin concentrations in P. aeruginosa with superior sensitivity and rapid turnaround. Getting to the final assay conditions for the LC/MS-based assay, we first implement a fluorescent assay, 2-(4-dimethylaminostyryl)-1-ethylpyridinium cation (DMP), and a radiolabeled version of the assay using 3H-labeled ciprofloxacin. Fluorescent probes, which change their spectroscopic

322

LC/MS-based drug accumulation assay/H. Cai et al./Anal.Biochem / H. Cai et al. / Anal. Biochem. 385 (2009) 321–325

properties in bacterial cytoplasm, are suitable for measuring uptake and efflux capabilities of bacteria. In general, these fluorescence probes are weakly fluorescent (short lifetimes) in aqueous environments and increase their quantum yield (long lifetimes) in nonpolar or hydrophobic environments. Using DMP, we established the conditions, including the cell density and plate type, for the assay. To translate the fluorescent assay to an LC/MS-based version, we used [3H]ciprofloxacin to evaluate the treatment time for the final assay format and the cell lysis procedure (required for the LC/MS-based assay) to ensure adequate compound recovery. Finally, assay reproducibility (inter- and intraday variabilities) was investigated using the established assay conditions.

Materials and methods Reagents DMP and carbonylcyanide m-chlorophenylhydrazone (CCCP) were purchased from Sigma. [3H]Ciprofloxacin and unlabeled ciprofloxacin were produced in-house by Pfizer. LB medium was purchased from Invitrogen. Bacterial strains and growth Strains used in this study were a generous gift from Herbert Schweizer. PAO1 (wild type), PAO200 (MexABOprM deficient mutant), and PAO314 (MexABOprM–MexCDOprJ–MexJKL:FRT deficient mutants) all were grown at 37 °C in LB medium. Experimental cultures were diluted 50-fold from overnight stocks and rotated at 200 rpm at 37 °C for 4 h. Optical density (OD) was measured at 600 nm. Preparation of cells for fluorescence measurement PAO1, PAO200, and PAO314 were grown at 37 °C overnight at approximately 225 rpm in LB medium. The next day, cells were harvested at 4000 rpm for 15 min, washed once with 100 mM NaCl–50 mM sodium phosphate buffer (pH 7.0), and resuspended in the same buffer to an OD600 of 1.5 in the presence of 0.05% glycerol. Assays were performed in 96-well plates (nontreated flat-bottom black plates from Costar) in a final volume of 200 ll for fluorescence measurement on a SpectraMax M2 or Fluorometric Imaging Plate Reader (FLIPR). Experimental measurements were generally performed within 2 h after cell suspension. Preparation of cells for radiometric measurement and LC/MS analysis PAO1 and PAO314 were grown at 37 °C overnight at approximately 225 rpm in LB medium. Experimental cultures were inoculated at a 1:50 dilution with a fully grown overnight culture and then rotated at approximately 225 rpm at 37 °C for approximately 4 h until an OD600 of approximately 1 was achieved. Cells were harvested by centrifugation at 3275 rpm for 15 min, washed once with 100 mM NaCl–50 mM sodium phosphate buffer (pH 7.0), and suspended again in the same buffer at 1:10 of the original volume in the presence of 0.5% glycerol and 1% glucose. Assays were performed in 96-well filter plates (Pall AcroPrep glass fiber) in a final volume of 200 ll. Experimental measurements were generally performed within 2 h after cell suspension. Fluorescence measurements Fluorescence measurements were performed at 30 °C on a SpectraMax M2 and at room temperature on a FLIPR. The fluorescent wavelength for DMP was 467/557 nm.

Radiometric and cold compound measurement Bacteria were preincubated with or without proton pump inhibitor CCCP (100 lM, used to abolish all of the efflux activities) for 10 min at 0 or 37 °C, and then [3H]ciprofloxacin or nonlabeled ciprofloxacin at varying concentrations was added for 10 min of uptake at either 0 or 37 °C. Blanks with no cells were included to account for nonspecific binding to the assay plate. After uptake time was complete, the cells were washed manually three times with ice-cold assay buffer using a Millipore vacuum manifold and vacu-pipette. Cells were broken open overnight with a 50-ll/ well volume of 100 mM glycine prepared in 0.1 N HCl. The flowthrough was transferred to a receiver plate by centrifugation at 3275g for 5 min. A 25-ll volume of the flow-through was then added to a LumaPlate and counted in a TopCount (PerkinElmer) scintillation counter (radiolabeled samples) or processed for LC/ MS analysis (nonlabeled assay). LC/MS analysis Sample and standard curve preparation was achieved by shaking 25 ll of sample or standard, 50 ll of water, and 20 ll of internal standard (240 ng/ml levofloxcin in water) and then injecting 3 to 4 ll into LC/tandem mass spectrometry (MS/MS). Quantitation was accomplished by comparison with a ciprofloxacin standard curve of 0.24 to 250 ng/ml. The LC/MS/MS quantitation employed a Sciex 4000 in electrospray ionization (ESI) mode, a CTC Autosampler, and Shimadzu pumps and a mobile phase of 11% (90:10 acetonitrile/0.1% formic) and 89% (5:95 acetonitrile/0.1% formic), a flow rate of 0.25 ml/min isocratic, and a Varian Polaris C18-A column (5 lm, 50  2 mm) as well as an MG-2 2000 guard column. The following multiple reaction monitoring (MRM) transitions were monitored for the analytes of interest: ciprofloxacin (+) 332.3 ? 288.2, levofloxacin (internal standard) (+) 362.3 ? 318.0. Cellular concentration calculation in P. aeruginosa In the LC/MS-based assay, cells at 1 OD/ml were concentrated 10-fold and 100 ll of the concentrated cells was used in the accumulation assay. Thus, each well had approximately 1 OD worth of cells. Assuming that P. aeruginosa has 5  108 cells per OD and a cellular volume of 1 lm3 (based on literature information on Escherichia coli [13]), the total cellular volume of 1 OD P. aeruginosa is [5  108  1 lm3  5  108/1  1012 ml approximately 0.5 ll 3 3 (1 ml  1 cm  1000 mm  1  1012 lm3)  5  104 ml  0.5 ll] [14]. Therefore, the cellular concentration is equal to the measured concentration using LC/MS (in ng/ml) multiplied by the extraction volume (0.05 ml) and divided by the total cellular volume per well of 0.5 ll. For example, assuming that the measured ciprofloxacin concentration was 5 ng/ml, the cellular concentration would be 500 ng/ml [(5 ng/ml  0.05 ml)/0.5 ll  0.5 lg/ml  500 ng/ml]. Results Fluorescent bacterial penetration assay using DMP DMP acted as a probe to test and set up the initial assay conditions for the bacterial penetration/accumulation assay in P. aeruginosa. Three strains of P. aeruginosa were tested: wild type (PAO1), MexABOprM deletion (PAO200), and MexABOprM–MexCDOprJ– MexJKL:FRT deletion mutants (PAO314). First, different types of 96-well plates were compared (data not shown here), and the flat-bottom black plate (nontreated from Costar) showed the lowest background noise and was chosen for further experiments.

323

Secondly, two cell densities were evaluated. Results from these studies demonstrated that at both cell densities the addition of CCCP resulted in an increased accumulation of DMP in wild-type cells (Fig. 1). Furthermore, the accumulation of DMP in pump knockout strains was similar with or without DMP (Fig. 1). Together, these results demonstrate that DMP is a substrate of OprM and CCCP affects the function of OprM, consistent with previous studies. Radiometric bacterial penetration assay using [3H]ciprofloxacin Time course study of compound accumulation was performed with 30 ng/ml [3H]ciprofloxacin at 0 and 37 °C. The amount of accumulation was calculated by subtracting accumulation at 0 °C from that at 37 °C, and the final results are graphed in Fig. 2 (raw data in Table 1). These experiments showed a continued accumulation of ciprofloxacin in the cell up to 30 min, followed by a decrease at the 1-h point in all cases. As with DMP, the addition of CCCP or the deletion of efflux pump in wild-type strains resulted in an increased intracellular accumulation. To adopt the assay for LC/MS analysis, an additional cell lysis step is required for release of compound (out of the cells) prior to quantitation using LC/MS. Side-by-side experiments with or without cell lysis treatment (with glycine–HCl or formic acid) using [3H]ciprofloxacin were performed to assess the compound recovery, and the results are tabulated in Table 2. LC/MS-based bacterial penetration assay using cold ciprofloxacin Assay reproducibility of the LC/MS-based assay using the final lysis condition was evaluated by assessing inter- and intraday assay variabilities, and the data are plotted in Fig. 3.

A

70000

no treatment

60000

treated w/ CCCP

RFU

40000 30000 20000 10000 0

B

70000 60000

PAO200

PAO314

no treatment treated w/ CCCP

RFU

50000 40000 30000 20000 10000 0 PAO1

PAO200

7000 6000 5000 4000 3000 2000 1000 0

0

10

20

30 40 Time (min)

50

60

Fig. 2. Accumulation time courses of 30 ng/ml [3H]ciprofloxacin in wild-type and pump knockout P. aeruginosa: s, PAO1; h, PAO1 with CCCP treatment; 4, PAO314; e, PAO314 with CCCP treatment. Drug concentrations were chosen to be well below the ciprofloxacin MIC (64 ng/ml for strain PAO1 [11]) so that cell integrity and function are not disrupted during the time course of the study. CPM, counts per minute. All data are means ± SD (n = 4). [Note: all final numbers were corrected for accumulation of ciprofloxacin at 0 °C and nonspecific binding to assay plates (controls with no cells).]

Table 1 Accumulation time courses of 30 ng/ml [3H]ciprofloxacin in wild-type and pump knockout P. aeruginosa

PAO1 PAO314 PAO1 with CCCP PAO314 with CCCP

5 min

15 min

30 min

60 min

178 ± 11 1431 ± 110 1542 ± 152 2041 ± 200

2271 ± 205 5018 ± 380 4974 ± 375 5271 ± 372

3713 ± 281 5635 ± 401 5123 ± 411 5616 ± 382

1111 ± 78 4866 ± 397 3809 ± 384 5096 ± 413

Note. Values are in counts per minute (CPM).

Discussion

50000

PAO1

Ciprofloxacin accumulation (CPM)

LC/MS-based drug accumulation assay/H. Cai et al./Anal.Biochem / H. Cai et al. / Anal. Biochem. 385 (2009) 321–325

PAO314

Fig. 1. Accumulation of 25 lM DMP in wild-type and pump knockout P. aeruginosa for 12.5 min at room temperature. The proton gradient was collapsed by treating P. aeruginosa with 100 lM CCCP for at least 10 min. (A) Cell density of OD 1. (B) Cell density of OD 1.5. RFU, relative fluorescence units. All data are means ± SD (n = 4 or 8).

To eventually set up an LC/MS-based bacterial accumulation assay for P. aeruginosa, a step-by-step approach was undertaken. A fluorescent probe, DMP, was first chosen to evaluate the pump mutants and validate the initial assay conditions (plate format and cell density) because it provided fast read-out, convenience, and minimal manipulation to the cells. As expected, at steady state the order of DMP accumulation was PAO314 > PAO200 > PAO1 (Fig. 1) because PAO314 had the most efflux pump deletion followed by PAO200. In addition, pretreatment with 100 lM CCCP, a proton gradient inhibitor, prior to the addition of DMP increased the fluorescence of wild-type cells to the levels achieved in PAO314 and was consistent with the hypothesis that the proton gradient was the driving force for the efflux pumps. Furthermore, two cell densities (ODs = 1 and 1.5) were evaluated (Fig. 1). At both densities, DMP accumulated at least twofold higher in pump knockout PAO314 or CCCP treatment groups than in wild-type PAO1. However, higher cell density did not offer any statistically significant gain in differentiation between wild-type and pump mutants (compared with a lower density of OD = 1) and, therefore, was not used in the subsequent experiments. The fluorescent bacterial penetration assay was a one-step format where the fluorescent measurement was done in situ without any cell isolation or manipulation after the addition of DMP. For an LC/MS-based assay, this approach will need to be modified accordingly; for example, cell separation and lysis procedures will need to be added. Thus, the second step we took was to conduct a radiolabeled assay using [3H]ciprofloxacin that resembled an LC/ MS-based assay format with the addition of a cell separation step. Because ciprofloxacin is a potent inhibitor against P. aeruginosa with minimal inhibitory concentration (MIC) numbers of 0.06 lg/

324

LC/MS-based drug accumulation assay/H. Cai et al./Anal.Biochem / H. Cai et al. / Anal. Biochem. 385 (2009) 321–325

Table 2 Assessment of [3H]ciprofloxacin with overnight glycine–HCl or formic acid treatment

Without glycine With glycine With formic acid

PAO1

PAO1 with CCCP

PAO314

PAO314 with CCCP

393 ± 23 404 ± 104 300 ± 64

1681 ± 248 1450 ± 24 761 ± 198

1813 ± 135 2086 ± 416 957 ± 148

2253 ± 248 2671 ± 311 1306 ± 359

Note. Values are in counts per minute (CPM). All final numbers were corrected for accumulation of ciprofloxacin in P. aeruginosa at 0 °C and nonspecific binding to assay plates (controls with no cells).

250 200

experiment 1

A

experiment 2 experiment 3

150

Cellular ciprofloxacin conc (ng/ml)

100 50 0 PAO1 -50

1000 800

PAO1 with CCCP

PAO314

experiment 1

PAO314 with CCCP

B

experiment 2 experiment 3

600 400 200 0 PAO1

PAO1 with CCCP

PAO314

PAO314 with CCCP

Fig. 3. Accumulation of 25 and 100 ng/ml ciprofloxacin in wild-type and efflux pump knockout P. aeruginosa for 10 min as measured using the LC/MS-based assay. (A) [Ciprofloxacin] = 25 ng/ml. (B) [Ciprofloxacin] = 100 ng/ml. Inter- and intraday variabilities were evaluated. All data are means ± SD (n = 4) [Note: all final numbers were corrected for accumulation of ciprofloxacin at 0 °C and nonspecific binding to assay plates (controls with no cells).].

ml (PAO1) and 0.006 lg/ml (PAO314) [15], we chose to use very small amounts of compound to minimize effects associated with antibacterial activity. In addition, two types of controls were added so as to (i) ensure minimal nonspecific binding to assay plates (blanks with no cells) and (ii) account for nonspecific binding to bacterial cell wall (accumulation measurement at 0 °C). At 30 ng/ ml, accumulation/uptake of [3H]ciprofloxacin was linear up to 30 min in wild-type PAO1 and 15 min in efflux mutant PAO314 and CCCP-treated bacteria. At 60 min, we observed a decrease in compound accumulation in PAO1 and only a slight decrease in the others. The underlying mechanism of the decreased accumulation in PAO1 is not currently understood; however, a treatment time of 10 min was chosen for the final step of assay validation to minimize antibacterial effects associated with ciprofloxacin treatment near the MIC. Treatment with glycine–HCl (pH 3.0) overnight to lyse the bacteria has been reported in the literature [5]. For our purpose, we evaluated two different lysis agents, formic acid and glycine–HCl, and found out that only glycine–HCl treatment provided near identical radiolabel counts relative to direct measurements of the

incubations. However, we needed to use a lower pH (pH 1.0) than what has been reported in literature (pH 3.0). Once the conditions were fully established, an important component of the assay validation is to assess the assay reproducibility by evaluating inter- and intraday assay variabilities. Preliminary experiments with a relatively high ciprofloxacin concentration (250 ng/ml) resulted in high background and was not employed in subsequent experiments. Based on the results from 25- and 100-ng/ml ciprofloxacin treatments (Fig. 3), approximately two- to threefold less accumulation (within standard errors) was consistently seen in PAO1 relative to PAO314 and CCCP-treated cells (Fig. 2) even through the absolute numbers were variable among experiments. One reason for such variability was that the background accumulation of the two controls was corrected; this could affect the final numbers but not the fold of differences. Therefore, it is recommended to use the absolute numbers (or fold of differences) for SARs within the same experiment but to use the fold of differences instead when data are coming from multiple runs. Furthermore, there are two points in Fig. 3 that are worth highlighting. First, the higher ciprofloxacin concentration (100 ng/ml) apparently did not generate greater differentiation between wildtype and pump knockout strains, as the data indicated. Thus, when applying the methodology, the preferred tested concentration for any compound would be low assuming that there is no issue in detecting a trace to small amount of compound in the accumulation study. Second, based on the approximation, the intracellular concentrations of ciprofloxacin were found to be higher in wildtype P. aeruginosa (PAO1) (up to twofold) and in PAO1 treated with CCCP and PAO314 (ranging from four- to sixfold) for 25- and 100ng/ml treatment groups (compared with the external concentrations in the medium). These observations are consistent with literature reports on tetracyclines and fluoroquinolones [16]. Tetracyclines are thought to reach an intracellular equilibrium concentration approximately twice that in medium because of the pH gradient between external medium and cytoplasm [17]. Similarly, fluoroquinolones are hypothesized to achieve high cytoplasmic concentration at equilibrium because quinolones chelate Mg2+ and the resulting positively charged compound may preferentially diffuse through porin channels. This could ultimately produce an elevated accumulation in the periplasm relative to that in the external medium because of the interior-negative Donnan potential across the outer membrane [16,18,19]. In conclusion, we have developed and validated an LC/MSbased bacterial penetration methodology that can be easily adapted for any bacteria of interest (e.g., E. coli) and drug discovery program to aid in developing SARs around physicochemical properties that favor bacterial penetration and efflux avoidance independent of whole cell inhibitory activities. Based on our experience with ciprofloxacin, the LC/MS detection method provides exquisite sensitivity (0.24 ng/ml). When coupled with advanced in vitro cell-free methodologies to assess target inhibition kinetics, including target residence time (compound off-rates), the application of this methodology provides a much more complete assessment of all the parameters required to create a new antibiotic. Finally, this technology has the potential to be applied

LC/MS-based drug accumulation assay/H. Cai et al./Anal.Biochem / H. Cai et al. / Anal. Biochem. 385 (2009) 321–325

to pharmacokinetic (PK)/pharmacodynamic (PD) modeling. Whereas current practice uses plasma drug concentrations, this methodology assesses intracellular drug concentrations (in the bacteria) that can be used in the PK/PD modeling because they more closely represent the true drug concentrations at the targeted mechanism. Acknowledgments The authors acknowledge John Hollembaek for providing the MS support for ciprofloxacin and Vincent E. Groppi for engaging in helpful discussions during the process of assay development. References [1] D.J. Payne, M.N. Gwynn, D.J. Holmes, D.L. Pompliano, Drugs for bad bugs: confronting the challenges of antibacterial discovery, Nat. Rev. Drug Discov. 6 (2007) 29–40. [2] R.A. Copeland, D.L. Pompliano, T.D. Meek, Drug-target residence time and its implications for lead optimization, Nat. Rev. Drug Discov. 5 (2006) 730–739. [3] V. Koronakis, J. Eswaran, C. Hughes, Structure and function of TolC: the bacterial exit duct of proteins and drugs, Annu. Rev. Biochem. 73 (2004) 467– 489. [4] D.G. Thanassi, S.J. Hultgren, Multiple pathways allow protein secretion across the bacterial outer membrane, Curr. Opin. Cell Biol. 12 (2000) 420–430. [5] L.J. Piddock, M.M. Johnson, Accumulation of 10 fluoroquinolones by wild-type or efflux mutant Streptococcus pneumoniae, Antimicrob. Agents Chemother. 46 (2002) 813–820. [6] P.G.S. Mortimer, L.J.V. Piddock, A comparison of methods used for measuring the accumulation of quinolones by Enterobacteriaceae, Pseudomonas aeruginosa, and Staphylococcus aureus, J. Antimicrob. Chemother. 28 (1991) 639–653.

325

[7] X.-Z. Li, H. Nikaido, K. Poole, Role of MexA–MexB–OprM in antibiotic efflux Pseudomonas aeruginosa, Antimicrob. Agents Chemother. 39 (1995) 1948– 1953. [8] V. Ricci, L. Piddock, Accumulation of garenoxacin by Bacteroides fragilis compared to that of five fluroroquinolones, J. Antimicrob. Chemother. 52 (2003) 605–609. [9] M.J. Everett, Y.F. Jin, V. Ricci, L.J.V. Piddock, Contributions of individual mechanisms to fluoroquinolone resistance in 36 Escherichia coli strains isolated from humans and animals, Antimicrob. Agents Chemother. 40 (1996) 2380– 2386. [10] L.J.V. Piddock, Y.F. Jin, V. Ricci, A.E. Asuquo, Quinolones accumulation by Pseudomonas aeruginosa, Staphylococcus aureus, and Escherichia coli, J. Antimicrob. Chemother. 43 (1999) 61–70. [11] L.J.V. Piddock, V. Ricci, Accumulation of five fluoroquinolones by Mycobacterium tuberculosis H37Rv, J. Antimicrob. Chemother. 48 (2001) 787– 791. [12] V. Ricci, L.J.V. Piddock, Accumulation of norfloxacin by Bacterium fragilis, Antimicrob. Agents Chemother. 44 (2000) 2361–2366. [13] R. Smither, Use of a Coulter counter to detect discrete changes in cell numbers and volume during growth of Escherichia coli, J. Appl. Bacteriol. 39 (1975) 157– 165. [14] H.E. Kubitschek, J.A. Friske, Determination of bacterial cell volume with the Coulter counter, J. Bacteriol. 168 (1986) 1466–1467. [15] R. Chuanchuen, C.T. Narasaki, H.P. Schweizer, The MexJK efflux pump of Pseudomonas aeruginosa requires OprM for antibiotic efflux but not for efflux of triclosan, J. Bacteriol. 184 (2002) 5036–5044. [16] H. Nikaido, D.G. Thanassi, Penetration of lipophilic agents with multiple protonation sites into bacteria cells: tetracyclines and fluoroquinolones as examples, Antimicrob. Agents Chemother. 37 (1993) 1393–1399. [17] A. Yamaguchi, H. Ohmori, M. Kaneko-Ohdera, T. Nomura, T. Sawai, DpHdependent accumulation of tetracycline in Escherichia coli, Antimicrob. Agents Chemother. 35 (1991) 53–56. [18] H. Nikaido, M. Vaara, Molecular basis of bacterial outer membrane permeability, Microbiol. Rev. 49 (1985) 1–32. [19] J.B. Stock, B. Rauch, S. Roseman, Periplasmic space in Salmonella typhimurium and Escherichia coli, J. Biol. Chem. 252 (1977) 7850–7861.