Journal of Antimicrobial Chemotherapy (1991) 28, 639-653
A comparison of methods used for measuring the accumulation of quinolones by Enterobacteriaceae, Pseudomonas aeruginosa and Staphylococcus aureus P. G. S. Mortimer and L. J. V. Piddock
Accumulation of norfloxacin by Escherichia coli was studied with a range of published procedures that used either radioactively-labelled norfloxacin ("C and 3H) or the natural fluorescence of the quinolone for detection. All methods except bioassay generated comparable data. A method involving the detection of fluorescence was found to be the method of choice. This method was used to study the accumulation kinetics of riprofloxatin, lomefloxacin, fleroxacin, norfloxacin, and enoxacin by several species of Gram-negative bacteria, and a Staphylococcus aureus strain. Saturation and efflux kinetics were also studied. There was no saturation at a concentration of norfloxacin < SO mg/L. Norfloxacin efflux was minimal during the uptake assay as the samples were withdrawn into ice-cold buffer; however, when the cells were sampled into buffer at 37°C, up to 50% of cell-associated quinolone effluxed within 5 min.
Introduction
Quinolones arc broad spectrum synthetic antimicrobial agents (Wolfson & Hooper, 1985) whose primary mechanism of action probably involves inhibition of the intracellular enzyme DNA gyrase (Gellert, 1981), although alternative mechanisms have suggested that quinolones bind to specific sites on DNA created by DNA gyrase rather than to DNA gyrase itself (Shen et al., 1989). The ability of quinolones to permeate bacteria to their intracellular target is, therefore, an important factor in determining both spectrum and activity, and reducing penetration can decrease bacterial susceptibility to these drugs. Quinolones have several permeability barriers to overcome in order to obtain access to their target site. In Gram-negative bacteria, both the outer membrane and the cytoplasmic membrane have to be traversed, while in Gram-positive bacteria only one membrane has to be crossed. By studying mutants lacking particular outer membrane proteins (Omps), quinolones have been shown to penetrate the outer membrane of Gram-negative organisms via porins (Hirai et al., 1986a). This has been confirmed by the1 isolation of quinolone-resistant strains that lack OmpF (Hirai et al., 19866). Mutants lacking OmpF were less susceptible (two to four-fold) to all quinolones, whereas mutants deficient in OmpC were as susceptible as the wild-type strain. This is •Corresponding author. 0305-7453/91/100639+15 $02.00/0
639 © 1991 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Antimicrobial Agents Research Group, Department of Medical Microbiology, The University of Birmingham, Birmingham B15 2TJ, UK
640
P. G. S. Mortimer and L. J. V. PMdock i
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
thought to be due to the difference in the diameters of the two porins (OmpF, 1-2 nm; OmpC, 1 nm), with OmpC being too small for quinolone penetration (Hirai et al., 19866). The passage of quinolones through the outer membrane is not, however, limited to porins, and other penetration methods have been suggested (Chapman & Georgopapadakou, 1988), including the 'self-promoted' mechanism described previously for aminoglycosidc (Hancock, 1984). This mechanism involves the displacement of divalent cations that bridge adjacent lipopolysaccharide (LPS) molecules, thereby disrupting the structure of the outer membrane and exposing areas of the lipid bilayer (Martin & Beveridge, 1986). For a given quinolone, the contributions of the porin and non-porin pathways to total accumulation are dependent upon the hydrophobicity of the quinolone. The more hydrophilic the drug, the less able it is to penetrate through the phospholipid bilayer (Chapman & Georgopapadakou, 1988). The mechanism of transport across the cytoplasmic membrane is less clear, although various inhibitors have been used by several groups to study this process (Hirai et al., 19866; Bedard, Wong & Bryan, 1987; Kotera et al., 1987; Chapman & Georgopapadakou, 1988; Hooper et al., 1989; Diver, Piddock & Wise, 1990). A comparison of the data and conclusions of these studies is difficult because of differences in procedures and quinolones used. At present, the precise mechanism that quinolones use to traverse the cytoplasmic membrane is unknown. It may be an energyindependent passive diffusion process, as proposed by Bedard et al. (1987), Cohen et al. (1988) and Chapman & Georgopapadakou (1988), or an energy-dependent active transport mechanism, as postulated by Kotera et al. (1987) and Diver et al. (1990). Several procedures have been used to investigate quinolone accumulation by bacteria. These can be grouped into three main categories, based on the method employed for detecting accumulated quinolone (radiolabel, bioassay or fluorescence). Several groups (Bedard et al., 1987; Hooper et al., 1989; Diver et al., 1990) have used radiolabelled quinolones, followed by removal of samples at timed intervals, filtration and scintillation counting. The method used by Bedard et al. (1987) involved an additional centrifugation step following sample removal before filtration. Other techniques used for measuring quinolone accumulation have involved the use of nonradiolabelled compounds. Thus, the method of Hirai et al. (1986a) involved withdrawing 10 mL samples of mid-exponential phase bacteria exposed to quinolone, followed by centrifugation, resuspension of the cells in 1 mL saline and boiling for 7 min. Accumulated quinolone was then detected with a microbiological plate assay. The method of Kotera et al. (1989) was similar, except that HPLC was used to detect the quinolone. The method used by Chapman & Georgopapadakou (1988) is unusual in that it detects the quinolone by measuring natural fluorescence of the quinolone nucleus. The difficulties encountered in comparing the published data are increased by the variations in media, growth conditions and strains used by each group. Most groups have used Escherichia coli K12 derivatives (Hirai et al., 1986a; Bedard et al., 1987; Hooper et al., 1989; Diver et al., 1990), but Kotera et al. (1989) studied only clinical isolates. This study was designed to compare the published methods. E. coli K12 strains were used to standardize the methods combined with a concentration of norfloxacin/L for comparative studies. Accumulation of norfloxacin at this concentration has been studied by several groups, although Hooper et al. (1989) used 0-04 mg/L. Accumulation of other quinolones at a range of concentrations by E. coli and other species was also examined with a modified method for detection of fluorescence.
Qnfoolone •ccnmnlatioa
641
Table L Strains used in the study Strain number
E. coli E. coli E. coli E. coli E.coli E.coli E. coli E.coli E. coli E. cloacae S. marcescens K. pneumoniae P. aeruginosa S. aureus
KL16 AB1157 1114 PCO479 PCO415 PC2912 PC2909 AB2470 SC13 Al BIS H43 Gl F77
Phenotype/Genotype Hfr thi-1 recA SpoT wt wt, NCTC 10538 PhoE" OmpF + OmpC+ PhoE" OmpF" OmpC" PhoE" OmpF+ OmpC" PhoE+ OmpF" OmpC" recB21 recA wt, NCTC 10005 wt,NCTC 10211 wt, NCTC 9633 wt, NCTC 10662 wt, NCTC 8532
Source/Reference Smith (1984) Casaregola et al. (1982) PHLS, London, UK Korteland et al. (1982) Korteland et al. (1982) Korteland et al. (1982) Korteland et al. (1982) Casaregola et al. (1982) Casaregola et al. (1982) PHLS, London, UK PHLS, London, UK PHLS, London, UK PHLS, London, UK PHLS, London, UK
wt, wild-type.
Materials and methods Bacterial strains The strains used are listed in Table I.
Bacteriological media
These comprised the following: M9 minimal salts (Gibco), supplemented with glucose 02% w/v, casamino acids 0-5% w/v (Difco) and thiamine 5 mg/L (BDH); Iso-Sensitest broth and agar (Unipath); Nutrient broth no. 2 (Lab M); Nutrient agar (Unipath). All broth cultures for uptake assays were shaken. Antibiotics
These were obtained as gifts from the respective manufacturers: [l4C]-norfloxacin (6-62 /iCi/mg) and [piperazine-U-3H]-norfloxacin (44-6 mCi/mg) from Merck Sharp and Dohme Laboratories; fleroxacin from Roche Products; nalidixic acid from Sterling-Winthrop; enoxacin from Warner Lambert/Parke Davis; unlabelled norfloxacin from Merck Sharp and Dohme; lomefloxacin from Searle; ciprofloxacin from Bayer. All antibiotics were made up and stored according to the manufacturer's instructions. Determination of antibiotic susceptibility
Susceptibility testing was performed with a standard agar doubling dilution method, Iso-Sensitest agar, an inoculum of 10* cfu/spot, and incubation at 37°C overnight. The MIC was defined as the concentration at which no more than ten colonies were detected.
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Species
642
P. G. S. Mortimer and L. J. V. Ptddock
Measurement of norfloxacin accumulation
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
(i) Method A was essentially as described by Hooper et al., (1989, personal communication). Bacteria were grown at 37°C to an A33O of 075, harvested by ccntrifugation, washed in M9 medium, and resuspended in fresh M9 medium to a concentration equivalent to an A,*, of 90. pHJ-Norfloxacin (specific activity adjusted to 3-5 mCi/mg with unlabelled norfloxacin) was used at a concentration of 004 mg/L, or 16 mg/L for comparison with other techniques in this study. At timed intervals, 40 uL samples were removed and collected on Whatman glass microfibre filters (045 /xm) pre-soaked with M9 medium. Filters were washed immediately with 6 mL of M9 medium at 25°C, dried and scintillation-counted. Cell protein was determined by the method of Lowry et al. (1951). The data were expressed either as ng norfloxacin accumulated/mg cell protein, or ng norfloxacin/mg dry cell weight (ii) Method B was essentially as described by Bedard et al. (1987) for [14C]-enoxacin; however, in this study, [l4C]-norfloxacin was used so that methods could be compared. Bacteria were grown in nutrient broth to an A^, of 04, and [MC]-norfloxacin added to a final concentration of 16 mg/L. One mL samples were removed at timed intervals, centrifuged, resuspended in nutrient broth, filtered through prewashed filters (Pall Ultipore N66 Nylon 66 filters, 045 /mi) and scintillation-counted. The data were expressed as ng norfloxacin/mg cells. (iii) Method C was essentially as described by Diver et al. (1990). Bacteria were grown to an A^, of 02-03 in Iso-Sensitest broth, and [l4C]-norfloxacin added to a final concentration of 16 mg/L. Samples (05 mL) were removed at timed intervals, diluted 1:40 with phosphate buffered saline (PBS) at 25°C, and filtered immediately through nylon filters (Pall Ultipor N66 Nylon 66filters;045 fan). The filters were washed with sterile PBS at 25°C, dried and scintillation-counted. The data were expressed as ng norfloxacin/mg cells. (iv) Method D was essentially as described by Hirai et al. (1986a). Bacteria were grown to mid-exponential phase at 37°C in Iso-Sensitest broth, and norfloxacin added to a final concentration of either 10 or 16 mg/L. At 1 min intervals, 10 mL was removed, chilled on ice for no longer than 2 min and the cells harvested by centrifugation at 3000 g for 10 min. The cell pellet was washed with 2 mL PBS, resuspended in 1 mL PBS, and boiled for 7 min. Norfloxacin was not inactivated by boiling (data not shown). The concentration of norfloxacin was determined with a microbiological plate assay. (iv) Method E was essentially as described by Chapman & Georgopapadakou (1988). Bacteria were grown in Iso-Sensitest broth to an A ^ of 04-08, harvested by centrifugation (5600 g for 5 min at 25°Q, washed once with 50 mM sodium phosphate buffer (pH 7-0) at 4°C and resuspended in the same buffer at a concentration equivalent to an A,,,! of 20. Fleroxacin was added to a final concentration of 16 mg/L. Samples (05 mg/L) were removed at timed intervals, diluted by the addition of 2 ml 50 mM sodium phosphate buffer (pH 7-0), centrifuged (5600 g for 1 min), and then treated with 2 ml of 01 M glycine hydrochloride (pH 3-0) for 60 min at 25°C. The samples were then centrifuged at 5600 g for 5 min, and the fluorescence of the supernatant determined at the excitation and emission spectra described by Chapman & Georgopapadakou (1988) with a fluorescence spectrophotometer (Fluorolog 2; Spex Industries.Inc., Metochen, NJ, USA). (vi) Method F was a modification of method E. Bacteria were grown in Iso-Sensitest
Qtrioolooe sccmnnktioa
643
Effect of buffer temperature on apparent steady-state concentration Using method F, E. coli AB1157, and norfloxacin 16 mg/L, duplicate samples were withdrawn at each time interval. One set was treated on ice, while the other set was diluted with sodium phosphate buffer at 25°C. The data were expressed as ng norfloxacin/mg cells. Efflux of norfloxacin from E. coli AB1157 Cells were grown, washed in 50 mM sodium phosphate buffer (pH 7-0) and resuspended as for method F. Norfloxacin was added to a final concentration of 16 mg/L, and the cells incubated at 37°C for 5 min. A 500 nL sample wasremovedand the concentration of cell-associated norfloxacin was determined. The remaining cell suspension was centrifuged (3003 g for 10 min at 4°C), and then resuspended in fresh Iso-Sensitest broth (37°Q at a concentration equivalent to an A^,, of 20. Further 500 fiL samples were removed at 20 sec intervals into ice-cold 50 mM sodium phosphate buffer (pH 7-0), and the concentrations of norfloxacin in the cell pellet and supernatant after centrifugation were determined. Cells were washed as for method F, and the concentration of norfloxacin within the wash was also determined. All data were expressed as ng norfloxacin/mg cells. Results MICs All the strains tested were susceptible to quinolones (Table II). Accumulation of norfloxacin: comparison of methods The total time period of the assay in each procedure varied from 30 min (methods A, C and D) to 100 min (methods B and E), so initial comparative experiments were performed over a period of 60 min. The sampling time also varied, but was usually at 5 min intervals (methods A, B, D and E). Preliminary studies in this laboratory with
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
broth to an A ^ of 0-6-0-8, harvested by centrifugation (3003 g at 4°C for 10 min), washed and then resuspended as for method E. Quinolone was added and 0-5 mL samples removed at 10 sec intervals for 1 min, and then at 30 sec intervals for up to S min. The samples were diluted immediately into 1 mL chilled sodium phosphate buffer (pH 7-0) on ice, and then centrifuged immediately in a microcentrifuge at 12,000 g, preferably at 4°C, for 5 min. The cells were washed with 1 mL chilled sodium phosphate buffer (pH 7-0), and rccentrifuged for 5 min. The cell pellet was then treated with 1 mL glycine hydrochloride (pH 3-0) for 2 h (or overnight) at room temperature, and the samples centrifuged (12,000 g for 10 min). The supernatant was removed and the pellet discarded. The decanted supernatant was centrifuged for a further 5 min at 12,000 g. Thefluorescenceof the final supernatant was then determined at the relevant excitation and emission spectra (Chapman & Georgopapadakou, 1988) and compared with a standard curve of 0O1-10 fig quinolone/mL in 0-1 M glycine hydrochloride (pH 3-0). The excitation and emission wavelength for enoxacin and ciprofloxacin were 346 nm/411 nm and 279 nm/447 nm respectively.
644
P. G. S. Mortimer and L. J. V. Piddock Table IL MICs (mg/L) of quinolones for the strains studied (106 cfu) MICs
Species
Strain
E. E. E. E. S. K. P. S.
NCTC 10538 AB1157 KL16 NCTC 10005 NCTC 10211 NCTC 9633 NCTC 10662 NCTC 8532
coli coli coli cloacae marctscens pneumoniae aeruginosa aureus
NAL
FLX
CIP
ENX
NOR
2 4 2 4 8 512
012 0-12 012 012 025 025 2
0015 012 O03 O06 O06 O06 05
512
8
05
012 05 025 006 006 006 2 2
012 006 012 006 025 025 1 4
8
assays of up to 1 h, and by Hooper et al. (1989), showed that most quinolone was accumulated ( > 90%) within the first 5 min, by which time an essentially steady-state concentration was reached. Therefore, in this study, all subsequent procedures were investigated with sampling at 30 sec intervals for up to 5 min. Essentially, E. coli KL16 and AB1157 had identical accumulation kinetics for norfloxacin (data not shown). With method A, accumulation of norfloxacin was studied at concentrations of 0-04 and 16 mg/L to enable a comparison with the published data of Hooper et al. (1989) and other workers. At 16 mg/L, accumulation was rapid and occurred within 60-100 sec (Figure 1), after which time a steady state concentration of approximately 140 ng norfloxacin/mg cells was attained. The steady-state did not increase further with time. At 0-04 mg/L, a steady-state concentration of 0-12 ng norfloxacin/mg cells (0-9 ng norfloxacin/mg cell protein) was attained after 60 sec (data not shown). With method B, and a concentration of norfloxacin 16 mg/L, a steady-state concentration of 150 ng norfloxacin/mg dry cells was reached within 120 sec (Figure 1). The steady-state concentration did not increase after prolonged exposure for up to 60 min. The centrifugation step made it impossible to take samples at 10 sec intervals (as with the other procedures); therefore intervals of 1 min were used. The steady-state concentration observed was higher than that observed with the other methods. With method C, accumulation was measured at a range of quinolone concentrations. A concentration of 1 mg/L was the lowest used because of the low specific activity of the [14C]-norfloxacin. With norfloxacin 16 mg/L, norfloxacin HOng/mg cells was accumulated within 120 sec; this increased by 40 ng during the following 55 min. The kinetics of accumulation were comparable with those observed by method F at all concentrations studied (1, 5, 10 and 16 mg/L; Figure 1 and data not shown). With method D (bioassay) the data obtained were highly variable when compared with previous studies (Table III). A prerequisite of any microbiological plate assay is the availability of a sensitive, reliable organism that gives good zone sizes at low concentrations of antibiotic. Two derivatives of E. coli AB1157 were used that are hypersusceptible to quinolones (AB2470 and SCI3; Casaregola, D'Ari & Huisman, 1982) with MICs of < 0-03 mg/L with all fluoroquinolones tested. With a concentration of norfloxacin 16 mg/L and AB2470 as the test organism, the apparent steadystate concentration varied from 150-500 ng norfloxacin /mg cells. With enoxacin 16 mg/L, the steady state attained after 10 min varied from 375-800 ng norfloxacin/mg
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
•NAL, Nalidixic acid; FLX,fleroxacin;CIP, ciprofloMcin, ENX, enoxacin; NOR, norfloxacin.
Qninolone accumulation
645
200
100
ISO 200 Time (tec)
300
250
350
Figure 1. Accumulation of norfloxacin (16 mg/L) by E. coJi AB1157, as measured by method A ( • ) , B, (A), C ( • ) and F (O).
cells with either AB2470 or SCI 3. Unsatisfactory data for nalidixic acid and ciprofloxadn were also obtained with SCI3 as the test organism. Initial problems of reproducibility with method E (Chapman & Georgopapadakou, 1988) were overcome by including additional centrifugation washing steps as detailed in method F. The accumulation of norfloxacin was studied with method F at a range of concentrations (0-1 mg/L to 50 mg/L). The steady-state concentration was proportional to the concentration of qninolone in the uptake assay (Table TV). Accumulation at all concentrations was rapid, with a steady-state concentration achieved within 60-120 sec. The lowest external concentration of quinolone that could be used in this procedure was 0-1 mg/L. At concentrations of > 50 mg norfloxacin /L accumulation was reduced and rapid killing was seen (data not shown). To determine the similarity between the data derived with each method, correlation coefficients were calculated by linear regression. Data from method F compared well with the data from methods A and C, with correlation coefficients of 0-85 and 0-95 Table HI. Data for quinolone accumulation obtained with a bioassay as the method of detection (method D)
Reference
Quinolone studied
Concentration (mg/L)
Duration of assay (min)
Steady state concentration (ng/mg cells)
Bedard et a!. (1987) Hirai et al. (1986a) This study This study
enoxacin norfloxacin enoxacin norfloxacin
20 10 20 16
10 10 10 10
1000 400
375-800 150-500
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
50
646
P. G. S. Mortimer and L. J. V. PkMock T«We IV. Relationship between the external concentration of norfloxacin and the steady-state concentration attained in E. coli as assayed by method F
External norfloxacin concentration (mg/L)
0 1-564 8-989 37126 74-7 97-0 1160 261-0
I n all experiment* the steady-ctau concentration was achieved within 60 tec.
respectively. Methods A, B and C also compared well with each other, with correlation coefficients of 0-86 (A VJ Q and 0-89 (B vs A). The results obtained with method F were reproducible; thus for the three E. coli strains studied (KL16, AB1157 and NCTC 10538; Table I), the mean standard deviation in the data obtained from five assays at all sampling times was 12 ng norfloxacin /mg cells, while for the same strains the mean standard deviation between assays was lower. The effect of temperature on method F The accumulation norfloxacin (at a concentration of 16 mg/L) by E. coli AB1157 was monitored with the standard protocol, except that the diluting buffer was at either 4° or 25°C. Samples diluted into ice-cold buffer reached an apparent steady state concentration of approximately 120 ng norfloxacin /mg cells within 120 sec, while the samples removed into buffer at 25°C reached a lower apparent steady-state concentration of 50 ng norfloxacin /mg cells after 30 sec (Figure 2). Efflux of norfloxacin from E. coli Cells were incubated with 16 mg norfloxacin /L for 5 min, collected by centrifugation, resuspended in fresh broth and efflux monitored. After 6 min there was a reduction of ~ 50% in cell-associated norfloxacin (Figure 3). Most norfloxacin effluxed within the first 3 min, after which no further efflux occurred. The norfloxacin concentration in each wash was also noted; approximately 4 ng norfloxacin /mg cells was detected, showing that little norfloxacin was lost during the washing of the cells in ice-cold sodium phosphate buffer. Accumulation of norfloxacin by E. coli mutants with altered expression ofporin proteins Accumulation of norfloxacin (at a concentration of 10 mg/L) was investigated with method F and defined porin-dencient mutants of E. coli (Table I and Figure 4). The
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
0 0-1 1 5 10 15 20 50
Steady-state concentration after 5 min' (ng/mg cells)
Qnfawlom accamaktion
647
140
60 60
steady-state concentration of the wildtype E. coli PCO479 was 79 ng norfloxacin/mg cells after 120 sec, while an E. coli strain lacking OmpC (PC 2912) attained a lower concentration of 60 ng norfloxacin/mg cells. E. coli PCO415, an ompR mutant (OmpF~ and OmpC"), had a Significantly lower steady-state concentration of 27 ng
0
20 40 60 80 100 120 140 160 ISO 200 220 240 260 280 300 320
Tim* (tec) Figure 3. The rate of norfloxacin efflux from E. coli ABl 157 after treatment with norfloxadn 16 mg/L, as detennined from the concentration of norfloxadn within the cells ( # ) , the nwtium ( • ) , and that lost during the washing stages (A)-
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
100 120 140 160 180 200 220 240 260 280 300 Tinw(MC) FIgBrc 2. The effect of temperature on the apparent accumulation of norfloxadn (at a concentration of 16 mg/L) by E. coli ABl 157, as determined by method F with buffer at 0°C ( • ) or 25°C ( • ) . 20 40
P. G. S. Mortimer and L. J. V. Piddock
648 100
40
60 80 100 120 140 160 180 200 220 240 260 280 300 Time (MC) Figure 4. Accumulation of norfloxacin (10 mg/L), at determined with method F, by porin-deficient mutants of E. coli: PCO415, PhoE- OmpF- OmpF- (A), PC2909, PhoE+ OmpF" OmpC" (O); PC2912, PhoE" OmpF+ OmpC" ( • ) ; PC0479, PhoE" OmpF+ OmpC+ (#).
norfloxacin/mg cells. Despite the loss of OmpF and OmpC when PhoE was produced constitutively (PC2909), the accumulation observed was similar to that seen for an OmpF + OmpC" PhoE" strain (PC2912). With all mutants, steady-state concentrations were attained after 120-180 sec. Accumulation of norfloxacin by Gram-negative and Gram-positive bacteria Accumulation by E. coli AB1157, Staphylococcus aureus NCTC 8532, Enterobacter cloacae NCTC 10005, Pseudomonas aeruginosa NCTC 10662, Serratia marcescens NCTC 10211 and Klebsiella pneumoniae NCTC 9633 was studied with norfloxacin at concentrations of either 16 or 10 mg/L. With norfloxacin 16 mg/L, E. cloacae, S. marcescens and S. aureus had similar accumulation profiles to E. coli AB1157, reaching steady-state concentrations of 140, 120 and 110 ng norfloxacin/mg cells respectively. With norfloxacin 10 mg/L, the K. pneumoniae strain, when compared with P. aeruginosa and 5". aureus had a lower steady-state concentration of 45 ng norfloxacin/mg cells (Figure 5). The time taken to reach a steady-state concentration was similar for all strains ( ~ 120 sec) except S. aureus ( > 240 sec). Initial problems in measuring accumulation by P. aeruginosa were overcome by growing the culture to a lower cell density. This prevented the cell pellets becoming glutinous and ensured that all cell-bound quinolone was released for detection. Accumulation of other quinolones as determined by method F The accumulation of ciprofloxacin (10 mg/L) was studied for several species. A lower steady-state was reached compared to norfloxacin; however, the time taken to attain a
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
20
Qutnokme accumulation
649
80
.
•
60 -A
B #
>^^
•••'
•
g 40
/
^
•
"
"K1/ i
i
20
40
i
i
60 80
i
i
i
i
i
i
i
i
i
I
i
100 120
140 160 180 200 220 240 260 280 300 Titnt (MC) Figure 5. Accumulation of norfloxatin (10 mg/L), u determined with method F by: P. aentginosa (A); K. pnewnoniae (#); S. aureus ( • ) .
steady-state was shorter (60-90 sec; Table V). With fleroxacin (16 mg/L), a steady-state concentration of 60 ng fleroxacdn/mg cells was attained within 180 sec by E. coli NCTC 10538, while for 5. aureus NCTC 8532 a steady-state concentration of 90 ng fleroxacin/mg cells was achieved after 120 sec. Accumulation of lomefloxacin was investigated with E. coli KL-16, S. aureus NCTC 8532, at concentrations of 1, 5, 10 and 16 mg/L (Piddock, Hall & Wise, 1990). Accumulation of enoxacin and nalidixic acid was studied with E. coli AB1157, S. aureus NCTC 8532 and P. aeruginosa NCTC 10662, but the results obtained were highly variable. With nalidixic acid 10 mg/L, the lack of a fluorinated group on the quinolone nucleus prevented its accurate detection by fluorescence spectrophotometry. Enoxacin is fluorinated, but the data obtained were unreliable for S. aureus, with an apparent six-fold increase in the steady-state concentration when a bacterial suspension was exposed to 16 mg/L compared with 10 mg/L (the steady-state concentrations being 180 and 28 ng enoxacin /mg cells respectively). Table V. The accumulation of ciprofloxacin (10 mg/L) by different species of bacteria
Species
Strain number
E.coli E. coli E. coli E. cloacae S. tnarcescens K. pnewnoniae P. aeruginosa S. aureus
NCTC 10538 AB 1157 KL16 NCTC 10005 NCTC 10211 NCTC 9633 NCTC 10662 NCTC 8532
Steady-state concentration (ng dprofloxacin/mg cells) 58 53 63 63 54 39 . 50 80
Time to reach the steady-state concentration (sec) 60-90 60-90 60 60-90 60 60
60 60-90
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
»
20 -
650
P. G. S. Mortimer and L. J. V. Ptddock
With enoxacin 10 mg/L and E. coli AB1157, a steady-state concentration of 60 ng enoxacin /mg cells was attained within 120 sec. Discussion
As an alternative to radiolabelling, method F (the revision of method E; Chapman & Georgopapadakou, 1988) was found to produce data comparable with those obtained by the procedures involving radiolabelled norfloxacin. Method F is a sensitive method which enables uptake to be detected with an external quinolone concentration of 0-1 mg/L. It has been suggested that the concentration of quinolone used to monitor accumulation should be < the MIC; however, accumulation was shown to be non-saturable up to an external concentration of 50 mg/L, and similar kinetics occurred at a range of concentrations. In addition, since > 90% of the steady-state concentration of quinolone enters bacterial cells within 5 min, and it is at least 10 min before significant inhibition of DNA synthesis occurs (Chow et al., 1988; Piddock, Walters & Diver, 1990), it is thought that measurements using quinolone 10 or 16 mg/L are significant and valid. Method F has several additional advantages over other methods in that it is less laborious, allows rapid sampling, and does not require the use of a radiolabelled drug. The only limitation is that a fluorescence spectrophotometer is required. Strict adherence to the assay protocol was important, since it was an essential requirement that samples were washed in ice-cold buffer and centrifuged immediately to differentiate between cell bound and accumulated quinolone. In comparison, the data obtained from samples added to buffer at room temperature showed marked differences in the time taken to reach steady-state and the steady-state concentration, suggesting that quinolone efflux is prevented by cooling the cells. The greatest efflux occurred within 2 min of sampling, and was presumed to be free norfloxacin which had effluxcd either across the cytoplasmic membrane into the periplasmic space, or from the
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Similar accumulation kinetics and steady-state concentrations were observed with the procedures including radiolabelled norfloxacin at a concentration of 16 mg/L, and these compared well with data published previously. The slightly higher steady-state seen with method A probably resulted from insufficient washing. With 0-04 mg/L, a steady-state concentration of ~ 012 ng norfloxacin/mg cells was achieved, equivalent to 0-9 ng norfloxacin/mg cell protein, which is comparable with the data of Hooper et al. (1989). Similar accumulation kinetics were observed with 0-04 and 16 mg/L, and a steady-state concentration was reached within 60 sec. Hooper et al. (1989) adjusted their data, so that they were expressed as ng norfloxacin/mg dry cells, by assuming that cellular protein was approximately 0-55 of cellular dry weight (Neidhart, 1987). In the present study the cellular protein of E. coli KL16, as determined by the method of Lowry et al. (1951), was 0-12 of cellular dry weight; with this factor the data from method A compared well with those obtained by the other methods. Method B (Bedard et al., 1987) gave slightly higher steady-state concentration values than methods A, C or F; however, the protocol does not involve any washing stages, although the authors recommend the addition of a wash step as it reduces the steady-state concentration. Bedard et al. (1987) studied enoxacin; however, with methods C and F, accumulation kinetics and steady-state concentrations for enoxacin and norfloxacin were similar. The major disadvantages of methods employing radiolabelled compounds are that labelled drugs are expensive, not easily available and require the use of classified areas for handling and disposal. In addition, as with all radiolabelled procedures, the sensitivity of the assay is dependent on the specific activity of the radiolabel.
Qtrfnokme •ccmnnlatioa
651
Reference? Ashby, J., Piddock, L. J. V. & Wise, R. (1985). An investigation of the hydrophobicity of the quinolones. Journal of Antimicrobial Chemotherapy 16, 805-8.
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
periplasm alone, and then diffused out of the cell via an unknown route across the outer membrane. Decreased temperature inhibited this apparent efflux, suggesting inhibition of a proposed active mechanism located at the cytoplasmic membrane (Cohen et al., 1988). Monitoring the washing stages of method F showed that some quinolone was lost, however, this was < 7% of the total and may represent norfloxacin diffusing from the periplasm, or bound to the cell surface. Method D (bioassay) generated the least reliable data, with a marked lack of rcproducibility. The technique was insensitive for several quinolones, laborious, and the mechanics of the process made quick sampling difficult A microbiological plate assay was also unreliable with a high level of inherent error (10%). Kotera et al. (1989) revised method D to include the use of HPLC as a detection, but this technique was not investigated in the present study. However, preparation of samples for HPLC requires several manipulations, such as deproteinization and ion-pair extraction, making this a laborious process compared with the other procedures, and especially method F. Studies of quinolone accumulation have frequently been performed over a 1 h period; however, this study and other published data indicate that > 90% of the steady-state is achieved within 5 min. In a study of the accumulation of 16 quinolones, only a few agents accumulated a further 5-10% of the concentration originally achieved after 5 min (A. D. Ebri Asuquo and L. J. V. Piddock, unpublished data). Accumulation of norfloxacin (10 mg/L) by E. coli mutants with altered expression of porin proteins was also studied. Loss of OmpC had little effect on accumulation, whereas loss of OmpF caused a reduction in the final steady-state concentration; however, the time taken to reach steady-state was the same as the corresponding OmpF + strain, thereby agreeing with the data of Hirai et al. (1986a). When PhoE was expressed in an ompR strain, accumulation kinetics similar to those of a strain lacking OmpC were observed, suggesting that, whilst OmpF is the major porin route for quinolones into E. coli, PhoE can also facilitate entry. The ability to use PhoE, but not OmpC, is unlikely to be due to pore diameter, which is similar for both porins ( ~ 11 nm), but may result from differences in the charge exposed on the inner surface of the pore channel. Studies of the accumulation of various quinolones by several different species showed that method F could be used with bacteria other than E. coli, although accumulation varied considerably from species to species. The final steady-state concentration, and the time required to reach the steady-state, differed according to the particular quinolone and species tested. Accumulation of all agents tested reached a maximum concentration within 180 sec, suggesting that the outer membrane is ineffective as a barrier. This is not unexpected as quinolones are hydrophilic compounds with partition coefficients of < 1 (Ashby, Piddock & Wise, 1985; Hirai et al., 1986A). The rate of diffusion of each drug is also thought to be influenced by the ionic charge (Chapman & Georgopapadakou, 1988), and this may account for the different values for different agents in the same strain. Overall, for wild-type bacteria, the steady-state concentration showed little relationship to susceptibility (e.g. E. coli and E. cloacae), and it would, therefore, appear that affinity to the target site governs this factor. For strains carrying mutations affecting the outer membrane, a reduced steady-state correlated with an increase in MIC.
652
P. G. S. Mortimer and L. J. V. Piddock
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Bcdard, J., Wong, S. & Bryan, L. E. (1987). Accumulation of enoxacin by Escherichia coli and Bacillus subtilis. Antimicrobial Agents and Chemotherapy 31, 1348-54. Casaregola, S., D'Ari, R. & Hirisman, O. (1982). Quantitative evaluation of recA gene expression in Escherichia coli. Molecular and General Genetics 185, 430-9. Chapman, J. S. & Georgopapadakou, N. H. (1988). Routes of quinolone permeation in Escherichia coli. Antimicrobial Agents and Chemotherapy 32, 438-42. Chow, R. T., Dougherty, T. J., Fraimow, H. S., Bellin, E. Y. & Miller, M. H. (1988). Association between early inhibition of DNA synthesis and the MICs and MBCs of carboxyquinolone antimicrobial agents for wild-type and mutant (gyrA nfxB (OmpF) acrA) Escherichia coli K-12. Antimicrobial Agents and Chemotherapy 32, 1113—8. Cohen, S. P., Hooper, D. C , Wolfson, J. S., Souza, K. S., McMurry, L. M. & Levy, S. B. (1988). Endogenous active efflux of norfloxacin in susceptible Escherichia coli. Antimicrobial Agents and Chemotherapy 32, 1187-91. Diver, J. M., Piddock, L. J. V. & Wise, R. (1990). The accumulation of five quinolone antibacterial agents by Escherichia coli. Journal of Antimicrobial Chemotherapy 25, 319-33. Gellert, M. (1981). DNA topoisomerases. Annual Review of Biochemistry 50, 879-910. Hancock, R. E. W. (1984). Alterations in outer membrane permeability. Annual Review of Microbiology 38, 237-64. Hirai, K., Aoyama, H., Irikura, T., Iyobe, S. & Mitsuhashi, S., (1986a). Differences in susceptibility to quinolones of outer membrane mutants of Salmonella typhimwium and Escherichia coli. Antimicrobial Agents and Chemotherapy 29, 535-8. Hirai, K., Aoyama, H., Suzue, S., Irikura, T., Iyobc, S. & Mitsuhashi, S. (1986*). Isolation and characterization of norfioxacin-resistant mutants of Escherichia coli K-12. Antimicrobial Agents and Chemotherapy 30, 248-53. Hooper, D. C , Wolfson, J. S., Souza, K. S., Ng, E. Y., McHugh, G. L. & Swartz, M. N. (1989). Mechanisms of quinolone resistance in Escherichia coli: characterization of nfx B and cfx B, two mutant resistance loci decreasing norfloxacin accumulation. Antimicrobial Agents and Chemotherapy 33, 283-90. Korteland, J., Tommassen, J. & Lugtenberg, B. (1982). PhoE protein pore of the outer membrane of Escherichia coli K.12 is a particularly efficient channel for organic and inorganic phosphate. Biochimica et Biophysica Acta 690, 282-9. Kotera, Y., Inoue, M. & Mitsuhashi, S. (1989). Potent antibacterial activity of KB-5246. In Program and Abstracts of the Twenty-Ninth Interscience Conference on Antimicrobial Agents and Chemotherapy. Houston. TX, 1989. Abstract 1242, p. 312. American Society for Microbiology, Washington, DC. Kotera, Y., Kato, T., Hirose, T., Inoue, M. & Mitsuhashi, S. (1987). Mechanism of norfloxacin uptake in Escherichia coli. In Progress in Antimicrobial and Anticancer Chemotherapy, Antimicrobial Section 2. Proceedings of the 15th International Congress of Chemotherapy, Istanbul. 1987 (Berkada, B. & Kuemmerle, H.-P., Eds), pp. 1909-11. Ecomed, Landsberg/Lech, Germany. Lowry, O. H., Rosenbrough, N. J., Farr, A. L. & Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry 193, 265-75. Martin, N. L. & Beveridge, T. J. (1986). Gentamkin interaction with Pseudomonas aeruginosa cell envelope. Antimicrobial Agents and Chemotherapy 29, 1079—87. Neidhardt, F. C. (1987). Chemical composition of Escherichia coli. In Escherichia coli and Salmonella typhimwium: Cellular and Molecular Biology, Vol. 1 (Neidhardt, F. C , Ed.), pp. 3-6. American Society for Microbiology, Washington, DC. Piddock, L. J. V., Hall, M. C. & Wise, R. (1990). Mechanism of action of lomefloxacin. Antimicrobial Agents and Chemotherapy 34, 1088-93. Piddock, L. J. V., Walters, R. & Diver, J. M. (1990). Correlation of quinolone MIC and inhibition of DNA, RNA, and protein synthesis and induction of the SOS response in Escherichia coli. Antimicrobial Agents and Chemotherapy 34, 2331-6. Shen, L. L., Kohlbrenner, W. E., Weigl, D. & Baranowski, J. (1989). Mechanism of quinolone inhibition of DNA gyrase. Appearance of unique norfloxacin binding sites in enzyme-DNA complexes. Journal of Biological Chemistry 264, 2973-8. Smith, J. T. (1984). Awakening the slumbering potential of the 4-quinolone antibacterials. Pharmaceutical Journal 233, 299-305.
Quinokme accumulation
653
Wolfson, J. S. & Hooper, D. C. (1985). The fluoroquinolones: structures, mechanisms of action and resistance, and spectra of activity in vitro. Antimicrobial Agents and Chemotherapy 28, 581-6. {Received 20 December 1990; revised version accepted 22 June 1991)
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
A comparison of methods used for measuring the accumulation of quinolones by Enterobacteriaceae, Pseudomonas aeruginosa and Staphylococcus aureus P. G. S. Mortimer and L. J. V. Piddock
Accumulation of norfloxacin by Escherichia coli was studied with a range of published procedures that used either radioactively-labelled norfloxacin ("C and 3H) or the natural fluorescence of the quinolone for detection. All methods except bioassay generated comparable data. A method involving the detection of fluorescence was found to be the method of choice. This method was used to study the accumulation kinetics of riprofloxatin, lomefloxacin, fleroxacin, norfloxacin, and enoxacin by several species of Gram-negative bacteria, and a Staphylococcus aureus strain. Saturation and efflux kinetics were also studied. There was no saturation at a concentration of norfloxacin < SO mg/L. Norfloxacin efflux was minimal during the uptake assay as the samples were withdrawn into ice-cold buffer; however, when the cells were sampled into buffer at 37°C, up to 50% of cell-associated quinolone effluxed within 5 min.
Introduction
Quinolones arc broad spectrum synthetic antimicrobial agents (Wolfson & Hooper, 1985) whose primary mechanism of action probably involves inhibition of the intracellular enzyme DNA gyrase (Gellert, 1981), although alternative mechanisms have suggested that quinolones bind to specific sites on DNA created by DNA gyrase rather than to DNA gyrase itself (Shen et al., 1989). The ability of quinolones to permeate bacteria to their intracellular target is, therefore, an important factor in determining both spectrum and activity, and reducing penetration can decrease bacterial susceptibility to these drugs. Quinolones have several permeability barriers to overcome in order to obtain access to their target site. In Gram-negative bacteria, both the outer membrane and the cytoplasmic membrane have to be traversed, while in Gram-positive bacteria only one membrane has to be crossed. By studying mutants lacking particular outer membrane proteins (Omps), quinolones have been shown to penetrate the outer membrane of Gram-negative organisms via porins (Hirai et al., 1986a). This has been confirmed by the1 isolation of quinolone-resistant strains that lack OmpF (Hirai et al., 19866). Mutants lacking OmpF were less susceptible (two to four-fold) to all quinolones, whereas mutants deficient in OmpC were as susceptible as the wild-type strain. This is •Corresponding author. 0305-7453/91/100639+15 $02.00/0
639 © 1991 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Antimicrobial Agents Research Group, Department of Medical Microbiology, The University of Birmingham, Birmingham B15 2TJ, UK
640
P. G. S. Mortimer and L. J. V. PMdock i
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
thought to be due to the difference in the diameters of the two porins (OmpF, 1-2 nm; OmpC, 1 nm), with OmpC being too small for quinolone penetration (Hirai et al., 19866). The passage of quinolones through the outer membrane is not, however, limited to porins, and other penetration methods have been suggested (Chapman & Georgopapadakou, 1988), including the 'self-promoted' mechanism described previously for aminoglycosidc (Hancock, 1984). This mechanism involves the displacement of divalent cations that bridge adjacent lipopolysaccharide (LPS) molecules, thereby disrupting the structure of the outer membrane and exposing areas of the lipid bilayer (Martin & Beveridge, 1986). For a given quinolone, the contributions of the porin and non-porin pathways to total accumulation are dependent upon the hydrophobicity of the quinolone. The more hydrophilic the drug, the less able it is to penetrate through the phospholipid bilayer (Chapman & Georgopapadakou, 1988). The mechanism of transport across the cytoplasmic membrane is less clear, although various inhibitors have been used by several groups to study this process (Hirai et al., 19866; Bedard, Wong & Bryan, 1987; Kotera et al., 1987; Chapman & Georgopapadakou, 1988; Hooper et al., 1989; Diver, Piddock & Wise, 1990). A comparison of the data and conclusions of these studies is difficult because of differences in procedures and quinolones used. At present, the precise mechanism that quinolones use to traverse the cytoplasmic membrane is unknown. It may be an energyindependent passive diffusion process, as proposed by Bedard et al. (1987), Cohen et al. (1988) and Chapman & Georgopapadakou (1988), or an energy-dependent active transport mechanism, as postulated by Kotera et al. (1987) and Diver et al. (1990). Several procedures have been used to investigate quinolone accumulation by bacteria. These can be grouped into three main categories, based on the method employed for detecting accumulated quinolone (radiolabel, bioassay or fluorescence). Several groups (Bedard et al., 1987; Hooper et al., 1989; Diver et al., 1990) have used radiolabelled quinolones, followed by removal of samples at timed intervals, filtration and scintillation counting. The method used by Bedard et al. (1987) involved an additional centrifugation step following sample removal before filtration. Other techniques used for measuring quinolone accumulation have involved the use of nonradiolabelled compounds. Thus, the method of Hirai et al. (1986a) involved withdrawing 10 mL samples of mid-exponential phase bacteria exposed to quinolone, followed by centrifugation, resuspension of the cells in 1 mL saline and boiling for 7 min. Accumulated quinolone was then detected with a microbiological plate assay. The method of Kotera et al. (1989) was similar, except that HPLC was used to detect the quinolone. The method used by Chapman & Georgopapadakou (1988) is unusual in that it detects the quinolone by measuring natural fluorescence of the quinolone nucleus. The difficulties encountered in comparing the published data are increased by the variations in media, growth conditions and strains used by each group. Most groups have used Escherichia coli K12 derivatives (Hirai et al., 1986a; Bedard et al., 1987; Hooper et al., 1989; Diver et al., 1990), but Kotera et al. (1989) studied only clinical isolates. This study was designed to compare the published methods. E. coli K12 strains were used to standardize the methods combined with a concentration of norfloxacin/L for comparative studies. Accumulation of norfloxacin at this concentration has been studied by several groups, although Hooper et al. (1989) used 0-04 mg/L. Accumulation of other quinolones at a range of concentrations by E. coli and other species was also examined with a modified method for detection of fluorescence.
Qnfoolone •ccnmnlatioa
641
Table L Strains used in the study Strain number
E. coli E. coli E. coli E. coli E.coli E.coli E. coli E.coli E. coli E. cloacae S. marcescens K. pneumoniae P. aeruginosa S. aureus
KL16 AB1157 1114 PCO479 PCO415 PC2912 PC2909 AB2470 SC13 Al BIS H43 Gl F77
Phenotype/Genotype Hfr thi-1 recA SpoT wt wt, NCTC 10538 PhoE" OmpF + OmpC+ PhoE" OmpF" OmpC" PhoE" OmpF+ OmpC" PhoE+ OmpF" OmpC" recB21 recA wt, NCTC 10005 wt,NCTC 10211 wt, NCTC 9633 wt, NCTC 10662 wt, NCTC 8532
Source/Reference Smith (1984) Casaregola et al. (1982) PHLS, London, UK Korteland et al. (1982) Korteland et al. (1982) Korteland et al. (1982) Korteland et al. (1982) Casaregola et al. (1982) Casaregola et al. (1982) PHLS, London, UK PHLS, London, UK PHLS, London, UK PHLS, London, UK PHLS, London, UK
wt, wild-type.
Materials and methods Bacterial strains The strains used are listed in Table I.
Bacteriological media
These comprised the following: M9 minimal salts (Gibco), supplemented with glucose 02% w/v, casamino acids 0-5% w/v (Difco) and thiamine 5 mg/L (BDH); Iso-Sensitest broth and agar (Unipath); Nutrient broth no. 2 (Lab M); Nutrient agar (Unipath). All broth cultures for uptake assays were shaken. Antibiotics
These were obtained as gifts from the respective manufacturers: [l4C]-norfloxacin (6-62 /iCi/mg) and [piperazine-U-3H]-norfloxacin (44-6 mCi/mg) from Merck Sharp and Dohme Laboratories; fleroxacin from Roche Products; nalidixic acid from Sterling-Winthrop; enoxacin from Warner Lambert/Parke Davis; unlabelled norfloxacin from Merck Sharp and Dohme; lomefloxacin from Searle; ciprofloxacin from Bayer. All antibiotics were made up and stored according to the manufacturer's instructions. Determination of antibiotic susceptibility
Susceptibility testing was performed with a standard agar doubling dilution method, Iso-Sensitest agar, an inoculum of 10* cfu/spot, and incubation at 37°C overnight. The MIC was defined as the concentration at which no more than ten colonies were detected.
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Species
642
P. G. S. Mortimer and L. J. V. Ptddock
Measurement of norfloxacin accumulation
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
(i) Method A was essentially as described by Hooper et al., (1989, personal communication). Bacteria were grown at 37°C to an A33O of 075, harvested by ccntrifugation, washed in M9 medium, and resuspended in fresh M9 medium to a concentration equivalent to an A,*, of 90. pHJ-Norfloxacin (specific activity adjusted to 3-5 mCi/mg with unlabelled norfloxacin) was used at a concentration of 004 mg/L, or 16 mg/L for comparison with other techniques in this study. At timed intervals, 40 uL samples were removed and collected on Whatman glass microfibre filters (045 /xm) pre-soaked with M9 medium. Filters were washed immediately with 6 mL of M9 medium at 25°C, dried and scintillation-counted. Cell protein was determined by the method of Lowry et al. (1951). The data were expressed either as ng norfloxacin accumulated/mg cell protein, or ng norfloxacin/mg dry cell weight (ii) Method B was essentially as described by Bedard et al. (1987) for [14C]-enoxacin; however, in this study, [l4C]-norfloxacin was used so that methods could be compared. Bacteria were grown in nutrient broth to an A^, of 04, and [MC]-norfloxacin added to a final concentration of 16 mg/L. One mL samples were removed at timed intervals, centrifuged, resuspended in nutrient broth, filtered through prewashed filters (Pall Ultipore N66 Nylon 66 filters, 045 /mi) and scintillation-counted. The data were expressed as ng norfloxacin/mg cells. (iii) Method C was essentially as described by Diver et al. (1990). Bacteria were grown to an A^, of 02-03 in Iso-Sensitest broth, and [l4C]-norfloxacin added to a final concentration of 16 mg/L. Samples (05 mL) were removed at timed intervals, diluted 1:40 with phosphate buffered saline (PBS) at 25°C, and filtered immediately through nylon filters (Pall Ultipor N66 Nylon 66filters;045 fan). The filters were washed with sterile PBS at 25°C, dried and scintillation-counted. The data were expressed as ng norfloxacin/mg cells. (iv) Method D was essentially as described by Hirai et al. (1986a). Bacteria were grown to mid-exponential phase at 37°C in Iso-Sensitest broth, and norfloxacin added to a final concentration of either 10 or 16 mg/L. At 1 min intervals, 10 mL was removed, chilled on ice for no longer than 2 min and the cells harvested by centrifugation at 3000 g for 10 min. The cell pellet was washed with 2 mL PBS, resuspended in 1 mL PBS, and boiled for 7 min. Norfloxacin was not inactivated by boiling (data not shown). The concentration of norfloxacin was determined with a microbiological plate assay. (iv) Method E was essentially as described by Chapman & Georgopapadakou (1988). Bacteria were grown in Iso-Sensitest broth to an A ^ of 04-08, harvested by centrifugation (5600 g for 5 min at 25°Q, washed once with 50 mM sodium phosphate buffer (pH 7-0) at 4°C and resuspended in the same buffer at a concentration equivalent to an A,,,! of 20. Fleroxacin was added to a final concentration of 16 mg/L. Samples (05 mg/L) were removed at timed intervals, diluted by the addition of 2 ml 50 mM sodium phosphate buffer (pH 7-0), centrifuged (5600 g for 1 min), and then treated with 2 ml of 01 M glycine hydrochloride (pH 3-0) for 60 min at 25°C. The samples were then centrifuged at 5600 g for 5 min, and the fluorescence of the supernatant determined at the excitation and emission spectra described by Chapman & Georgopapadakou (1988) with a fluorescence spectrophotometer (Fluorolog 2; Spex Industries.Inc., Metochen, NJ, USA). (vi) Method F was a modification of method E. Bacteria were grown in Iso-Sensitest
Qtrioolooe sccmnnktioa
643
Effect of buffer temperature on apparent steady-state concentration Using method F, E. coli AB1157, and norfloxacin 16 mg/L, duplicate samples were withdrawn at each time interval. One set was treated on ice, while the other set was diluted with sodium phosphate buffer at 25°C. The data were expressed as ng norfloxacin/mg cells. Efflux of norfloxacin from E. coli AB1157 Cells were grown, washed in 50 mM sodium phosphate buffer (pH 7-0) and resuspended as for method F. Norfloxacin was added to a final concentration of 16 mg/L, and the cells incubated at 37°C for 5 min. A 500 nL sample wasremovedand the concentration of cell-associated norfloxacin was determined. The remaining cell suspension was centrifuged (3003 g for 10 min at 4°C), and then resuspended in fresh Iso-Sensitest broth (37°Q at a concentration equivalent to an A^,, of 20. Further 500 fiL samples were removed at 20 sec intervals into ice-cold 50 mM sodium phosphate buffer (pH 7-0), and the concentrations of norfloxacin in the cell pellet and supernatant after centrifugation were determined. Cells were washed as for method F, and the concentration of norfloxacin within the wash was also determined. All data were expressed as ng norfloxacin/mg cells. Results MICs All the strains tested were susceptible to quinolones (Table II). Accumulation of norfloxacin: comparison of methods The total time period of the assay in each procedure varied from 30 min (methods A, C and D) to 100 min (methods B and E), so initial comparative experiments were performed over a period of 60 min. The sampling time also varied, but was usually at 5 min intervals (methods A, B, D and E). Preliminary studies in this laboratory with
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
broth to an A ^ of 0-6-0-8, harvested by centrifugation (3003 g at 4°C for 10 min), washed and then resuspended as for method E. Quinolone was added and 0-5 mL samples removed at 10 sec intervals for 1 min, and then at 30 sec intervals for up to S min. The samples were diluted immediately into 1 mL chilled sodium phosphate buffer (pH 7-0) on ice, and then centrifuged immediately in a microcentrifuge at 12,000 g, preferably at 4°C, for 5 min. The cells were washed with 1 mL chilled sodium phosphate buffer (pH 7-0), and rccentrifuged for 5 min. The cell pellet was then treated with 1 mL glycine hydrochloride (pH 3-0) for 2 h (or overnight) at room temperature, and the samples centrifuged (12,000 g for 10 min). The supernatant was removed and the pellet discarded. The decanted supernatant was centrifuged for a further 5 min at 12,000 g. Thefluorescenceof the final supernatant was then determined at the relevant excitation and emission spectra (Chapman & Georgopapadakou, 1988) and compared with a standard curve of 0O1-10 fig quinolone/mL in 0-1 M glycine hydrochloride (pH 3-0). The excitation and emission wavelength for enoxacin and ciprofloxacin were 346 nm/411 nm and 279 nm/447 nm respectively.
644
P. G. S. Mortimer and L. J. V. Piddock Table IL MICs (mg/L) of quinolones for the strains studied (106 cfu) MICs
Species
Strain
E. E. E. E. S. K. P. S.
NCTC 10538 AB1157 KL16 NCTC 10005 NCTC 10211 NCTC 9633 NCTC 10662 NCTC 8532
coli coli coli cloacae marctscens pneumoniae aeruginosa aureus
NAL
FLX
CIP
ENX
NOR
2 4 2 4 8 512
012 0-12 012 012 025 025 2
0015 012 O03 O06 O06 O06 05
512
8
05
012 05 025 006 006 006 2 2
012 006 012 006 025 025 1 4
8
assays of up to 1 h, and by Hooper et al. (1989), showed that most quinolone was accumulated ( > 90%) within the first 5 min, by which time an essentially steady-state concentration was reached. Therefore, in this study, all subsequent procedures were investigated with sampling at 30 sec intervals for up to 5 min. Essentially, E. coli KL16 and AB1157 had identical accumulation kinetics for norfloxacin (data not shown). With method A, accumulation of norfloxacin was studied at concentrations of 0-04 and 16 mg/L to enable a comparison with the published data of Hooper et al. (1989) and other workers. At 16 mg/L, accumulation was rapid and occurred within 60-100 sec (Figure 1), after which time a steady state concentration of approximately 140 ng norfloxacin/mg cells was attained. The steady-state did not increase further with time. At 0-04 mg/L, a steady-state concentration of 0-12 ng norfloxacin/mg cells (0-9 ng norfloxacin/mg cell protein) was attained after 60 sec (data not shown). With method B, and a concentration of norfloxacin 16 mg/L, a steady-state concentration of 150 ng norfloxacin/mg dry cells was reached within 120 sec (Figure 1). The steady-state concentration did not increase after prolonged exposure for up to 60 min. The centrifugation step made it impossible to take samples at 10 sec intervals (as with the other procedures); therefore intervals of 1 min were used. The steady-state concentration observed was higher than that observed with the other methods. With method C, accumulation was measured at a range of quinolone concentrations. A concentration of 1 mg/L was the lowest used because of the low specific activity of the [14C]-norfloxacin. With norfloxacin 16 mg/L, norfloxacin HOng/mg cells was accumulated within 120 sec; this increased by 40 ng during the following 55 min. The kinetics of accumulation were comparable with those observed by method F at all concentrations studied (1, 5, 10 and 16 mg/L; Figure 1 and data not shown). With method D (bioassay) the data obtained were highly variable when compared with previous studies (Table III). A prerequisite of any microbiological plate assay is the availability of a sensitive, reliable organism that gives good zone sizes at low concentrations of antibiotic. Two derivatives of E. coli AB1157 were used that are hypersusceptible to quinolones (AB2470 and SCI3; Casaregola, D'Ari & Huisman, 1982) with MICs of < 0-03 mg/L with all fluoroquinolones tested. With a concentration of norfloxacin 16 mg/L and AB2470 as the test organism, the apparent steadystate concentration varied from 150-500 ng norfloxacin /mg cells. With enoxacin 16 mg/L, the steady state attained after 10 min varied from 375-800 ng norfloxacin/mg
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
•NAL, Nalidixic acid; FLX,fleroxacin;CIP, ciprofloMcin, ENX, enoxacin; NOR, norfloxacin.
Qninolone accumulation
645
200
100
ISO 200 Time (tec)
300
250
350
Figure 1. Accumulation of norfloxacin (16 mg/L) by E. coJi AB1157, as measured by method A ( • ) , B, (A), C ( • ) and F (O).
cells with either AB2470 or SCI 3. Unsatisfactory data for nalidixic acid and ciprofloxadn were also obtained with SCI3 as the test organism. Initial problems of reproducibility with method E (Chapman & Georgopapadakou, 1988) were overcome by including additional centrifugation washing steps as detailed in method F. The accumulation of norfloxacin was studied with method F at a range of concentrations (0-1 mg/L to 50 mg/L). The steady-state concentration was proportional to the concentration of qninolone in the uptake assay (Table TV). Accumulation at all concentrations was rapid, with a steady-state concentration achieved within 60-120 sec. The lowest external concentration of quinolone that could be used in this procedure was 0-1 mg/L. At concentrations of > 50 mg norfloxacin /L accumulation was reduced and rapid killing was seen (data not shown). To determine the similarity between the data derived with each method, correlation coefficients were calculated by linear regression. Data from method F compared well with the data from methods A and C, with correlation coefficients of 0-85 and 0-95 Table HI. Data for quinolone accumulation obtained with a bioassay as the method of detection (method D)
Reference
Quinolone studied
Concentration (mg/L)
Duration of assay (min)
Steady state concentration (ng/mg cells)
Bedard et a!. (1987) Hirai et al. (1986a) This study This study
enoxacin norfloxacin enoxacin norfloxacin
20 10 20 16
10 10 10 10
1000 400
375-800 150-500
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
50
646
P. G. S. Mortimer and L. J. V. PkMock T«We IV. Relationship between the external concentration of norfloxacin and the steady-state concentration attained in E. coli as assayed by method F
External norfloxacin concentration (mg/L)
0 1-564 8-989 37126 74-7 97-0 1160 261-0
I n all experiment* the steady-ctau concentration was achieved within 60 tec.
respectively. Methods A, B and C also compared well with each other, with correlation coefficients of 0-86 (A VJ Q and 0-89 (B vs A). The results obtained with method F were reproducible; thus for the three E. coli strains studied (KL16, AB1157 and NCTC 10538; Table I), the mean standard deviation in the data obtained from five assays at all sampling times was 12 ng norfloxacin /mg cells, while for the same strains the mean standard deviation between assays was lower. The effect of temperature on method F The accumulation norfloxacin (at a concentration of 16 mg/L) by E. coli AB1157 was monitored with the standard protocol, except that the diluting buffer was at either 4° or 25°C. Samples diluted into ice-cold buffer reached an apparent steady state concentration of approximately 120 ng norfloxacin /mg cells within 120 sec, while the samples removed into buffer at 25°C reached a lower apparent steady-state concentration of 50 ng norfloxacin /mg cells after 30 sec (Figure 2). Efflux of norfloxacin from E. coli Cells were incubated with 16 mg norfloxacin /L for 5 min, collected by centrifugation, resuspended in fresh broth and efflux monitored. After 6 min there was a reduction of ~ 50% in cell-associated norfloxacin (Figure 3). Most norfloxacin effluxed within the first 3 min, after which no further efflux occurred. The norfloxacin concentration in each wash was also noted; approximately 4 ng norfloxacin /mg cells was detected, showing that little norfloxacin was lost during the washing of the cells in ice-cold sodium phosphate buffer. Accumulation of norfloxacin by E. coli mutants with altered expression ofporin proteins Accumulation of norfloxacin (at a concentration of 10 mg/L) was investigated with method F and defined porin-dencient mutants of E. coli (Table I and Figure 4). The
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
0 0-1 1 5 10 15 20 50
Steady-state concentration after 5 min' (ng/mg cells)
Qnfawlom accamaktion
647
140
60 60
steady-state concentration of the wildtype E. coli PCO479 was 79 ng norfloxacin/mg cells after 120 sec, while an E. coli strain lacking OmpC (PC 2912) attained a lower concentration of 60 ng norfloxacin/mg cells. E. coli PCO415, an ompR mutant (OmpF~ and OmpC"), had a Significantly lower steady-state concentration of 27 ng
0
20 40 60 80 100 120 140 160 ISO 200 220 240 260 280 300 320
Tim* (tec) Figure 3. The rate of norfloxacin efflux from E. coli ABl 157 after treatment with norfloxadn 16 mg/L, as detennined from the concentration of norfloxadn within the cells ( # ) , the nwtium ( • ) , and that lost during the washing stages (A)-
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
100 120 140 160 180 200 220 240 260 280 300 Tinw(MC) FIgBrc 2. The effect of temperature on the apparent accumulation of norfloxadn (at a concentration of 16 mg/L) by E. coli ABl 157, as determined by method F with buffer at 0°C ( • ) or 25°C ( • ) . 20 40
P. G. S. Mortimer and L. J. V. Piddock
648 100
40
60 80 100 120 140 160 180 200 220 240 260 280 300 Time (MC) Figure 4. Accumulation of norfloxacin (10 mg/L), at determined with method F, by porin-deficient mutants of E. coli: PCO415, PhoE- OmpF- OmpF- (A), PC2909, PhoE+ OmpF" OmpC" (O); PC2912, PhoE" OmpF+ OmpC" ( • ) ; PC0479, PhoE" OmpF+ OmpC+ (#).
norfloxacin/mg cells. Despite the loss of OmpF and OmpC when PhoE was produced constitutively (PC2909), the accumulation observed was similar to that seen for an OmpF + OmpC" PhoE" strain (PC2912). With all mutants, steady-state concentrations were attained after 120-180 sec. Accumulation of norfloxacin by Gram-negative and Gram-positive bacteria Accumulation by E. coli AB1157, Staphylococcus aureus NCTC 8532, Enterobacter cloacae NCTC 10005, Pseudomonas aeruginosa NCTC 10662, Serratia marcescens NCTC 10211 and Klebsiella pneumoniae NCTC 9633 was studied with norfloxacin at concentrations of either 16 or 10 mg/L. With norfloxacin 16 mg/L, E. cloacae, S. marcescens and S. aureus had similar accumulation profiles to E. coli AB1157, reaching steady-state concentrations of 140, 120 and 110 ng norfloxacin/mg cells respectively. With norfloxacin 10 mg/L, the K. pneumoniae strain, when compared with P. aeruginosa and 5". aureus had a lower steady-state concentration of 45 ng norfloxacin/mg cells (Figure 5). The time taken to reach a steady-state concentration was similar for all strains ( ~ 120 sec) except S. aureus ( > 240 sec). Initial problems in measuring accumulation by P. aeruginosa were overcome by growing the culture to a lower cell density. This prevented the cell pellets becoming glutinous and ensured that all cell-bound quinolone was released for detection. Accumulation of other quinolones as determined by method F The accumulation of ciprofloxacin (10 mg/L) was studied for several species. A lower steady-state was reached compared to norfloxacin; however, the time taken to attain a
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
20
Qutnokme accumulation
649
80
.
•
60 -A
B #
>^^
•••'
•
g 40
/
^
•
"
"K1/ i
i
20
40
i
i
60 80
i
i
i
i
i
i
i
i
i
I
i
100 120
140 160 180 200 220 240 260 280 300 Titnt (MC) Figure 5. Accumulation of norfloxatin (10 mg/L), u determined with method F by: P. aentginosa (A); K. pnewnoniae (#); S. aureus ( • ) .
steady-state was shorter (60-90 sec; Table V). With fleroxacin (16 mg/L), a steady-state concentration of 60 ng fleroxacdn/mg cells was attained within 180 sec by E. coli NCTC 10538, while for 5. aureus NCTC 8532 a steady-state concentration of 90 ng fleroxacin/mg cells was achieved after 120 sec. Accumulation of lomefloxacin was investigated with E. coli KL-16, S. aureus NCTC 8532, at concentrations of 1, 5, 10 and 16 mg/L (Piddock, Hall & Wise, 1990). Accumulation of enoxacin and nalidixic acid was studied with E. coli AB1157, S. aureus NCTC 8532 and P. aeruginosa NCTC 10662, but the results obtained were highly variable. With nalidixic acid 10 mg/L, the lack of a fluorinated group on the quinolone nucleus prevented its accurate detection by fluorescence spectrophotometry. Enoxacin is fluorinated, but the data obtained were unreliable for S. aureus, with an apparent six-fold increase in the steady-state concentration when a bacterial suspension was exposed to 16 mg/L compared with 10 mg/L (the steady-state concentrations being 180 and 28 ng enoxacin /mg cells respectively). Table V. The accumulation of ciprofloxacin (10 mg/L) by different species of bacteria
Species
Strain number
E.coli E. coli E. coli E. cloacae S. tnarcescens K. pnewnoniae P. aeruginosa S. aureus
NCTC 10538 AB 1157 KL16 NCTC 10005 NCTC 10211 NCTC 9633 NCTC 10662 NCTC 8532
Steady-state concentration (ng dprofloxacin/mg cells) 58 53 63 63 54 39 . 50 80
Time to reach the steady-state concentration (sec) 60-90 60-90 60 60-90 60 60
60 60-90
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
»
20 -
650
P. G. S. Mortimer and L. J. V. Ptddock
With enoxacin 10 mg/L and E. coli AB1157, a steady-state concentration of 60 ng enoxacin /mg cells was attained within 120 sec. Discussion
As an alternative to radiolabelling, method F (the revision of method E; Chapman & Georgopapadakou, 1988) was found to produce data comparable with those obtained by the procedures involving radiolabelled norfloxacin. Method F is a sensitive method which enables uptake to be detected with an external quinolone concentration of 0-1 mg/L. It has been suggested that the concentration of quinolone used to monitor accumulation should be < the MIC; however, accumulation was shown to be non-saturable up to an external concentration of 50 mg/L, and similar kinetics occurred at a range of concentrations. In addition, since > 90% of the steady-state concentration of quinolone enters bacterial cells within 5 min, and it is at least 10 min before significant inhibition of DNA synthesis occurs (Chow et al., 1988; Piddock, Walters & Diver, 1990), it is thought that measurements using quinolone 10 or 16 mg/L are significant and valid. Method F has several additional advantages over other methods in that it is less laborious, allows rapid sampling, and does not require the use of a radiolabelled drug. The only limitation is that a fluorescence spectrophotometer is required. Strict adherence to the assay protocol was important, since it was an essential requirement that samples were washed in ice-cold buffer and centrifuged immediately to differentiate between cell bound and accumulated quinolone. In comparison, the data obtained from samples added to buffer at room temperature showed marked differences in the time taken to reach steady-state and the steady-state concentration, suggesting that quinolone efflux is prevented by cooling the cells. The greatest efflux occurred within 2 min of sampling, and was presumed to be free norfloxacin which had effluxcd either across the cytoplasmic membrane into the periplasmic space, or from the
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Similar accumulation kinetics and steady-state concentrations were observed with the procedures including radiolabelled norfloxacin at a concentration of 16 mg/L, and these compared well with data published previously. The slightly higher steady-state seen with method A probably resulted from insufficient washing. With 0-04 mg/L, a steady-state concentration of ~ 012 ng norfloxacin/mg cells was achieved, equivalent to 0-9 ng norfloxacin/mg cell protein, which is comparable with the data of Hooper et al. (1989). Similar accumulation kinetics were observed with 0-04 and 16 mg/L, and a steady-state concentration was reached within 60 sec. Hooper et al. (1989) adjusted their data, so that they were expressed as ng norfloxacin/mg dry cells, by assuming that cellular protein was approximately 0-55 of cellular dry weight (Neidhart, 1987). In the present study the cellular protein of E. coli KL16, as determined by the method of Lowry et al. (1951), was 0-12 of cellular dry weight; with this factor the data from method A compared well with those obtained by the other methods. Method B (Bedard et al., 1987) gave slightly higher steady-state concentration values than methods A, C or F; however, the protocol does not involve any washing stages, although the authors recommend the addition of a wash step as it reduces the steady-state concentration. Bedard et al. (1987) studied enoxacin; however, with methods C and F, accumulation kinetics and steady-state concentrations for enoxacin and norfloxacin were similar. The major disadvantages of methods employing radiolabelled compounds are that labelled drugs are expensive, not easily available and require the use of classified areas for handling and disposal. In addition, as with all radiolabelled procedures, the sensitivity of the assay is dependent on the specific activity of the radiolabel.
Qtrfnokme •ccmnnlatioa
651
Reference? Ashby, J., Piddock, L. J. V. & Wise, R. (1985). An investigation of the hydrophobicity of the quinolones. Journal of Antimicrobial Chemotherapy 16, 805-8.
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
periplasm alone, and then diffused out of the cell via an unknown route across the outer membrane. Decreased temperature inhibited this apparent efflux, suggesting inhibition of a proposed active mechanism located at the cytoplasmic membrane (Cohen et al., 1988). Monitoring the washing stages of method F showed that some quinolone was lost, however, this was < 7% of the total and may represent norfloxacin diffusing from the periplasm, or bound to the cell surface. Method D (bioassay) generated the least reliable data, with a marked lack of rcproducibility. The technique was insensitive for several quinolones, laborious, and the mechanics of the process made quick sampling difficult A microbiological plate assay was also unreliable with a high level of inherent error (10%). Kotera et al. (1989) revised method D to include the use of HPLC as a detection, but this technique was not investigated in the present study. However, preparation of samples for HPLC requires several manipulations, such as deproteinization and ion-pair extraction, making this a laborious process compared with the other procedures, and especially method F. Studies of quinolone accumulation have frequently been performed over a 1 h period; however, this study and other published data indicate that > 90% of the steady-state is achieved within 5 min. In a study of the accumulation of 16 quinolones, only a few agents accumulated a further 5-10% of the concentration originally achieved after 5 min (A. D. Ebri Asuquo and L. J. V. Piddock, unpublished data). Accumulation of norfloxacin (10 mg/L) by E. coli mutants with altered expression of porin proteins was also studied. Loss of OmpC had little effect on accumulation, whereas loss of OmpF caused a reduction in the final steady-state concentration; however, the time taken to reach steady-state was the same as the corresponding OmpF + strain, thereby agreeing with the data of Hirai et al. (1986a). When PhoE was expressed in an ompR strain, accumulation kinetics similar to those of a strain lacking OmpC were observed, suggesting that, whilst OmpF is the major porin route for quinolones into E. coli, PhoE can also facilitate entry. The ability to use PhoE, but not OmpC, is unlikely to be due to pore diameter, which is similar for both porins ( ~ 11 nm), but may result from differences in the charge exposed on the inner surface of the pore channel. Studies of the accumulation of various quinolones by several different species showed that method F could be used with bacteria other than E. coli, although accumulation varied considerably from species to species. The final steady-state concentration, and the time required to reach the steady-state, differed according to the particular quinolone and species tested. Accumulation of all agents tested reached a maximum concentration within 180 sec, suggesting that the outer membrane is ineffective as a barrier. This is not unexpected as quinolones are hydrophilic compounds with partition coefficients of < 1 (Ashby, Piddock & Wise, 1985; Hirai et al., 1986A). The rate of diffusion of each drug is also thought to be influenced by the ionic charge (Chapman & Georgopapadakou, 1988), and this may account for the different values for different agents in the same strain. Overall, for wild-type bacteria, the steady-state concentration showed little relationship to susceptibility (e.g. E. coli and E. cloacae), and it would, therefore, appear that affinity to the target site governs this factor. For strains carrying mutations affecting the outer membrane, a reduced steady-state correlated with an increase in MIC.
652
P. G. S. Mortimer and L. J. V. Piddock
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
Bcdard, J., Wong, S. & Bryan, L. E. (1987). Accumulation of enoxacin by Escherichia coli and Bacillus subtilis. Antimicrobial Agents and Chemotherapy 31, 1348-54. Casaregola, S., D'Ari, R. & Hirisman, O. (1982). Quantitative evaluation of recA gene expression in Escherichia coli. Molecular and General Genetics 185, 430-9. Chapman, J. S. & Georgopapadakou, N. H. (1988). Routes of quinolone permeation in Escherichia coli. Antimicrobial Agents and Chemotherapy 32, 438-42. Chow, R. T., Dougherty, T. J., Fraimow, H. S., Bellin, E. Y. & Miller, M. H. (1988). Association between early inhibition of DNA synthesis and the MICs and MBCs of carboxyquinolone antimicrobial agents for wild-type and mutant (gyrA nfxB (OmpF) acrA) Escherichia coli K-12. Antimicrobial Agents and Chemotherapy 32, 1113—8. Cohen, S. P., Hooper, D. C , Wolfson, J. S., Souza, K. S., McMurry, L. M. & Levy, S. B. (1988). Endogenous active efflux of norfloxacin in susceptible Escherichia coli. Antimicrobial Agents and Chemotherapy 32, 1187-91. Diver, J. M., Piddock, L. J. V. & Wise, R. (1990). The accumulation of five quinolone antibacterial agents by Escherichia coli. Journal of Antimicrobial Chemotherapy 25, 319-33. Gellert, M. (1981). DNA topoisomerases. Annual Review of Biochemistry 50, 879-910. Hancock, R. E. W. (1984). Alterations in outer membrane permeability. Annual Review of Microbiology 38, 237-64. Hirai, K., Aoyama, H., Irikura, T., Iyobe, S. & Mitsuhashi, S., (1986a). Differences in susceptibility to quinolones of outer membrane mutants of Salmonella typhimwium and Escherichia coli. Antimicrobial Agents and Chemotherapy 29, 535-8. Hirai, K., Aoyama, H., Suzue, S., Irikura, T., Iyobc, S. & Mitsuhashi, S. (1986*). Isolation and characterization of norfioxacin-resistant mutants of Escherichia coli K-12. Antimicrobial Agents and Chemotherapy 30, 248-53. Hooper, D. C , Wolfson, J. S., Souza, K. S., Ng, E. Y., McHugh, G. L. & Swartz, M. N. (1989). Mechanisms of quinolone resistance in Escherichia coli: characterization of nfx B and cfx B, two mutant resistance loci decreasing norfloxacin accumulation. Antimicrobial Agents and Chemotherapy 33, 283-90. Korteland, J., Tommassen, J. & Lugtenberg, B. (1982). PhoE protein pore of the outer membrane of Escherichia coli K.12 is a particularly efficient channel for organic and inorganic phosphate. Biochimica et Biophysica Acta 690, 282-9. Kotera, Y., Inoue, M. & Mitsuhashi, S. (1989). Potent antibacterial activity of KB-5246. In Program and Abstracts of the Twenty-Ninth Interscience Conference on Antimicrobial Agents and Chemotherapy. Houston. TX, 1989. Abstract 1242, p. 312. American Society for Microbiology, Washington, DC. Kotera, Y., Kato, T., Hirose, T., Inoue, M. & Mitsuhashi, S. (1987). Mechanism of norfloxacin uptake in Escherichia coli. In Progress in Antimicrobial and Anticancer Chemotherapy, Antimicrobial Section 2. Proceedings of the 15th International Congress of Chemotherapy, Istanbul. 1987 (Berkada, B. & Kuemmerle, H.-P., Eds), pp. 1909-11. Ecomed, Landsberg/Lech, Germany. Lowry, O. H., Rosenbrough, N. J., Farr, A. L. & Randall, R. J. (1951). Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry 193, 265-75. Martin, N. L. & Beveridge, T. J. (1986). Gentamkin interaction with Pseudomonas aeruginosa cell envelope. Antimicrobial Agents and Chemotherapy 29, 1079—87. Neidhardt, F. C. (1987). Chemical composition of Escherichia coli. In Escherichia coli and Salmonella typhimwium: Cellular and Molecular Biology, Vol. 1 (Neidhardt, F. C , Ed.), pp. 3-6. American Society for Microbiology, Washington, DC. Piddock, L. J. V., Hall, M. C. & Wise, R. (1990). Mechanism of action of lomefloxacin. Antimicrobial Agents and Chemotherapy 34, 1088-93. Piddock, L. J. V., Walters, R. & Diver, J. M. (1990). Correlation of quinolone MIC and inhibition of DNA, RNA, and protein synthesis and induction of the SOS response in Escherichia coli. Antimicrobial Agents and Chemotherapy 34, 2331-6. Shen, L. L., Kohlbrenner, W. E., Weigl, D. & Baranowski, J. (1989). Mechanism of quinolone inhibition of DNA gyrase. Appearance of unique norfloxacin binding sites in enzyme-DNA complexes. Journal of Biological Chemistry 264, 2973-8. Smith, J. T. (1984). Awakening the slumbering potential of the 4-quinolone antibacterials. Pharmaceutical Journal 233, 299-305.
Quinokme accumulation
653
Wolfson, J. S. & Hooper, D. C. (1985). The fluoroquinolones: structures, mechanisms of action and resistance, and spectra of activity in vitro. Antimicrobial Agents and Chemotherapy 28, 581-6. {Received 20 December 1990; revised version accepted 22 June 1991)
Downloaded from http://jac.oxfordjournals.org/ at Georgetown University on March 9, 2015
