Journal of Antimicrobial Chemotherapy (1992) 29, 529-538
Postantibiotic effect of CI-960, enoxacin and dprofloxacin on Escherichia coir, effect on morphology and haemolysin activity
Department of Microbiology, Medical College of Ohio, Toledo, Ohio 43699, USA The postantibiotic effect (PAE) has been classically defined as the suppression of bacterial growth that persists after limited exposure of organisms to antimicrobial agents. Morphology and haemolysin activity during the PAE of three quinolones on Escherichia coli were examined in this study. A one hour exposure to the quinolones, CI-960, enoxacin and dprofloxacin, produced a PAE of 0-5-2-0 h. When determinated by Coulter counter, at 0-5 x MIC of enoxacin or CI-960 after 1 h exposure, 58% or 42% cells, respectively, of the treated cells were filamentous (cell length > 12 fan). After drug removal, the population of the filamentous cells decreased, however, after even 4 h, 12% and 2% of the cells were still filamentous after exposure to enoxacin or CI-960. Further morphological studies during the PAE showed that the first division of the filamentous cell was asymmetrical, and both bacterial cell division and septation were delayed after exposure to 0-5 MIC of CI-960. Following quinolone removal, the treated E. coli did not exhibit normal activity of haemolysin for at least 2 h. Internal haemolysin activity was adversely affected for 1 h. The results of this study suggest that any consideration of postantibiotic effects should include the residual antibiotic effects on bacterial morphology and virulence factors, in addition to the defined suppression of bacterial regrowth.
Introduction The postantibiotic effect (PAE) has been classically defined as a term used to describe the suppression of bacterial growth that persists after limited exposure of organisms to antimicrobial agents (Vogelman et al., 1988). The longer PAE, the lower the probability that growth of the infecting bacterium will resume during that fraction of the interdose interval where antimicrobial levels fall below the minimum inhibitory concentration (MIC) (Eliopoulos & Eliopoulos, 1989). A long PAE provides the potential for administering the antimicrobial agent at longer intervals between doses (Craig & Gudmunsson, 1986). This classically defined PAE usually only addresses the bacterial regrowth after a short exposure to an antimicrobial agent. Other bacterial physiological characters (especially virulence factors) during the PAE have not been well studied. If bacterial virulence factor production or activity decreased after exposure to antibiotics, the clinical relevance of PAE could be greater than previously considered. The primary target site for quinolones is DNA gyrase, which has been shown to be involved in a number of cellular processes other than supercoiling of bacterial DNA. DNA replication, transcription, and repair as well as recombination have been shown to be DNA-gyrase dependent (Hays & Bocchmer, 1978; Kreuzer & Cozzarclli, 1979). 529 0305-7453/92/050529 +10 $02.00/0
© 1992 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Iinong Gun and Jeffrey C. Burnham
530
L. Gnu and J. C Burnham
Materials and methods
Bacterial strains E. coli strains ATCC 25922 and J96 (O4, K6; provided by Sheila Hull, Baylor College of Medicine, Houston, TX), was isolated from a patient with acute pyelonephritis (Normark et al., 1983). Determination of MIC and MBC The MIC of CI-960 and enoxacin (Parke Davis Pharmaceutical Research Division, Warner-Lambert Company, Ann Arbor, MI), and ciprofloxacin and co-trimoxazole (Sigma Chemical Co., St. Louis, MO) for each strain were determined by microbroth dilution in Mueller-Hinton Broth (Difco). After determination of the approximate MIC, a more accurate MIC was determined by dividing the range between the 0-5 x MIC and 2 x MIC into ten dilutions. The MIC determinations were read after incubation at 37°C for 18-20 h. An inoculum of approximately 1 x 103 cfu/mL was used. The bacterial inoculum for MBC determination was prepared as above for the determination of the MICs. After 18-20 h of incubation with the drugs and 10- to 100fold dilution in sterile 0-85% N a d , bacteria were counted by subculturing 0-1 mL of bacteria directly onto Mueller-Hinton Agar (Difco) for evaluation after 18-20 h at 37°C. Experiments were performed in duplicate. The MBC was taken as the lowest concentration which gave 99-9% bacterial killing. PAE The reference tube macrodilution method was performed as described by Craig & Gudmunsson (1986). The inocula were prepared by incubating the bacteria overnight at 37°C in Mueller-Hinton broth. The culture was then diluted 1:100 in fresh prewarmed Mueller-Hinton broth and incubated in a shaking water bath until the
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Evidence has accumulated indicating that at higher doses, the quinolones and/or novobiocin interfere with transcription-translation of genes sensitive to catabolite repression but have no effect on constitutively expressed genes (Sanzey, 1979; Ostrowslri & Hulanicka, 1981; Dalhoff & Doring, 1985). The synthesis of virulence factors, e.g. proteases and exotoxin A in Pseudomonas aeruginosa was significantly reduced by subinhibitory concentrations of dprofloxatin under in-vitro as well as invivo conditions. The degree of inhibition of protease and toxin production was strainspecific but independent of metabolic inhibition (Dalhoff & Doring, 1985). Similar effects by quinolones at subinhibitory concentrations on virulence-associated factors in Escherichia coli have also been reported (Burnham & Sonstein, 1990). Chin & Neu (1987) showed that ciprofloxacin had a 3- to 6-h postantibiotic suppressive effect on growth of staphylococci, Enterobacteriaceae and P. aeruginosa, but not on Enterococcus faecalis. The aim of this work was to investigate and compare the PAEs of CI-960 (PD127391), a newer quinolone, along with two well studied quinolones, ciprofloxacin and enoxacin, with that of co-trimoxazole on E. coli. The response of bacterial morphology and haemolysin activity during the PAE period was also studied.
PAE of qninolones
531
Morphology The effects of antibiotics on bacterial morphology were determined by using a Coulter counter (Model ZM, Technical Communications, Coulter Electronics, Inc., USA). The population of the cells with abnormal size (cell length > 12 /an) were measured after exposure to antibiotics and subsequent drug removal. Bacterial cell division and septation were studied after exposure to 0-5 x MIC CI-960 for 1 h. The drug was removed by washing with fresh Mueller-Hinton broth and the washed cells were incubated on the surface of Mueller-Hinton agar. Bacterial morphological changes were monitored every 15 min by phase contrast microscopy (Axiophot, Carl Zeiss Instruments, Inc., Germany). Haemolysin activity Haemolysin activity was measured as described by Springer & Goebel (1980) with the modifications described as follows: Accumulation of haemolysin: 1 mL samples of E. coli cells were taken at intervals as described for the measurement of PAE. The cells were removed by centrifugation, and the supernatants assayed for haemolytic activity which was denned as 'external haemolysin'. The washed cell pellets were suspended in 1 mL of 0-01 M phosphate buffer (pH 6-0) and sonicated. The cell debris was removed . by centrifugation, and the supernatants were tested for haemolytic activity and defined as 'internal haemolysin'. Haemolytic activity was determined by the amount of haemoglobin released from washed erythrocytes: one hundred microlitres of external or internal haemolysin was added to 100 pL ox erythrocytes (buffered by pH 7-5, 0-01 M Tris-hydrochloride, 002 M calcium chloride) in microtitre plates. The assay plates were incubated at 40°C for 10 min, and incubation was terminated by quickly removing unlysed erythrocytes by centrifugation at 4°C. The amount of released haemoglobin was measured by the absorbance of the supernatants at 405 nm. One unit of haemolysin activity was defined as the amount of haemolysin protein which caused the release of 1 /anol of haemoglobin in 1 h under test conditions. Results MICs and MBCs The MICs and MBCs of each antibiotic for E. coli ATCC 25922 and J96 are given in Table I. CI-960 was the most active with the lowest MIC and MBC; the MBC being double the MIC. The MICs of ciprofloxacin and enoxacin were equal to the MBCs.
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
logarithmic phase of growth was reached. Cultures with a cell density of approximately 5x10* cfu/mL were then exposed to various drug concentrations for 1 h. In order to remove residual antibiotics completely, cultures were diluted 100- or 1000-fold in fresh pre-warmed Mueller-Hinton broth. Viable counts were determined on Mueller-Hinton agar plates for the initial inoculum, following the removal of the antibiotics. The PAE was calculated from the equation PAE (time) •= T-C, where T is the time required for the count in cfu/mL of the test culture to increase by one log10 above the count observed immediately after drug removal, and C is the time required for the count of cfu/mL in an untreated control culture to increase one log,0 above the count observed immediately after completion of the same procedure used on the culture for drug removal (Craig & Gudmunsson, 1986).
L. Guan and J. C Burnham
532
Table L Susceptibility to and PAE of E. coli exposed for 1 h to four antibiotics in different concentrations
MIC Strain ATCC 25922
(mg/L)
0004 0015
CI-960 ciprofloxacin enoxacin co-trimoxazole*
0*5
Q-960
O004
ciprofloxacin
003
enoxacin
019
co-trimoxazole
O60
0-16
Concentration (mg/L)
0016 (4 x MIQ 006 (4xMIQ 064 (4 x MIQ 2-40 (4 x MIQ 0016 (4 x MIQ O004 (1 x MIQ 0002 (05 x MIQ O001 (025 x MIQ 012 (4 x MIQ 003 (1 x MIQ 0015 (05 x MIQ O008 (025 x MIQ 076 (4 x MIQ 019 ( l x MIQ O095(O5xMIQ 0048 (025 x MIQ 2-40 (4 x MIQ O60 (1 x MIQ
030 (05 x MIQ
014 (025 x MIQ
PAE (h) 1-7 1-6
09 02 20 05 01 -01 1-8 1-2 09 05 1-4 1-2 07 02 02 00 00 ND
'1 trimethoprim: 5 lulfamethoxazole. ND, Not done.
PAE The PAE of the four drugs in different concentrations for the two strains of E. coli after exposure for 1 h is shown in Table I. A significant period of growth suppression was seen after exposure of E. coli ATCC 25922 and J96 to CI-960, ciprofloxacin and enoxacin at a concentration of 4 x MIC. For E. coli J96, CI-960 showed the longest PAE, 2-0 h, at 4 x MIC, and shortest PAE, 0-5 h, at 1 x MIC; whereas ciprofloxacin and enoxacin produced a PAE of 0-9 h and 0-7 h respectively, at 0-5 x MIC. CI-960 did not show significant PAE at sub-inhibitory concentrations. PAE was not seen with E. coli ATCC 25922 at sub-inhibitory concentrations with any of the quinolones. Co-trimoxazole did not show significant PAE at concentrations up to 4 x MIC with either strains. CI-960, ciprofloxacin and enoxacin were bactericidal at 4 x MIC for E. coli J96 after 1 h exposure (Table II). Enoxacin and ciprofloxacin also showed bacterial killing activity at 1 x MIC. All the drugs tested showed no bactericidal activity at concentrations below their MICs (Table II).
Morphology In order to avoid the interference from dead cells, quinolones were used at 0-5 x MIC for bacterial morphology and haemolysin activity assays.
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
J96
Antibiotic
533
PAE of qatnotonei
Table IL The viable count of E. coli J96 after one hour's exposure to CI-960, enoxadn and ciprofloxacin Antibiotic
Concentration (mg/L)
Viable cell count before exposure after exposure 8x10*
None CI-960
Ciprofloxacin
8xl07 2x10* 2x10'
99-75 0
8x10' 8x10' 2x10* 1x10* 2x10' 8x10'3
0
3X1O
0 97-50 87-00 0
0
4x10* 2x10'
99-96 50-00 0
8x10'
0
These morphological studies resulted in two major findings: (i) While the normal E. coli cell length is 1-3 fan (Wheat, 1988), filamentation of E. coli cells is a common phenomenon when exposed to quinolones (Diver & Wise, 1986). In this study the morphological changes of these filamentous cells after the antibiotic was removed were examined. Table i n shows the effects of 0-5 x MIC of CI-960 and enoxacin on E. coli J96 morphology as determined by Coulter counter. Only 0-12% of the population of antibiotic-free cells are longer than 12 fan. At 0-5 x MIC of enoxacin or CI-960, after exposure 1 h, 58% and 42% cells, respectively, were longer than 12 fan. After drug removal, the population of filamentous cells decreased. Even after 4 h, 12% of the enoxacin exposed cells were still filamentous, and 2% filamented of those treated with CI-960. These results show that the effects of CI-960 and enoxacin on E. coli morphology can last at least 4 h. Ciprofloxacin showed similar effects to enoxacin. Co-trimoxazole did not show significant effects at 0-5 x MIC. (ii) Figure 1 shows bacterial cell division and septation after exposure to 0-5 x MIC of CI-960. Cell 'X' was one of the treated cells. At time 0, just after washing, the length of cell X was 30 fan. There was no cell division observed after 105 min, although cell TaMe m. Effects of 0-5 x the MIC of CI-960 or enoxacin on E. coli J96 morphology (measured by Coulter counter) Time (h) 0* 1 2
3 4
Control CI-960 Enoxacin cells total cells cells total cells total cells cells counted > 12 urn (%) counted > 12 urn (%) counted > 12 um (%) 1-4X104 2-OxlO4 1-6X104 1-7 xlO 4 1-3 xlO 4
12 (0-12) 5(0-025) 0(0) 0(0)
0(0)
4-0 xlO 4 1-2 xlO 4 1-7 xlO 4 4-2 x 104
51 x 10*.
16810 (42)
6000(50) 3061 (18)
3362 (8) 102(2)
5-3 xlO 3 7-6 xlO 3 4-2 xlO 4 9-2 xlO 4 4-4x10*
3074 (58) 3648 (42) 15690 (38) 1840(20) 5280 (12)
Time 0 - after 1 b exposure to 0-5 x MIC of 0-960 or enoxacin and subsequent drug removal.
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Enoxacin
(K)16(4xMIQ 0O04 (1 x MIC) 0-002 (0-5 x MIC) 0O01 (0-25 x MIC) 0-76 (4 x MIC) 0-19 l x M I Q (H)95 (0-5 x MIC) (MM8 (0-25 x MIC) 0-12 (4 x MIC) 003 (1 x MIC) 0-015 (0-5 x MIC) 0-008 (0-25 x MIC)
Killing (%)
I
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
(c)
Figure 1. Light micrograph of E. coli J96 cell division and septation after exposure to 0-5 x the MIC of CI-960: (a) cell X at 0 min, PAE, i.e., after 1 h exposure and subsequent washing with Mueller-Hinton broth to remove the drug; (b) at 105 min; (c) at 120 min, cell XI and X2 are daughter cells of cell X; (d) at 135 min, cell XIA and XIB are daughter cells of cell XI; (e) at 150 min; (f) at 210 min, cell XlAl and X1A2 are daughter cells of cell XIA. Cell length: 1 cm = 7-5 fan.
PAE of qriookraes
535
Haemolysin To compare the haemolysin activity between the antibiotic-free control and antibiotictreated cells, cell mass (e.g., 'per O D ^ ' (unit of optical density at 520 nm)) was used as a reference unit Figure 2 shows the effects of 0-5 x MIC of enoxacin and cotrimoxazole on both external and internal haemolysin activity in E. coli J96 after removal of drug. The external haemolysin activity of 0-5 x MIC of enoxacin treated cells was inhibited for at least 2 h, about 1 h longer than PAE (the PAE was 0-7 h, Table I). One hour and 2 h after drug removal, the external haemolysin activity of the treated cells was about 56% and 65% of the activity of control cells (Figure 2(a)). The internal haemolysin activity of treated cells was strongly inhibited for 1 h. The activity of internal haemolysin of the treated cells was about 31 % of the control cells after drug removal at 1 h (Figure 2(c)). Ciprofloxacin showed similar effects to enoxacin on both external and internal haemolysin activity. Co-trimoxazole did not show any significant effect on the haemolysin activity of E. coli (Figure 2(b), (d)).
Discussion This study shows that CI-960, ciprofloxacin and enoxacin can produce significant PAE on E. coli. An increase in the concentration of the three quinolones was associated with prolongation of PAE. This is similar to data reported by other workers (Craig & Gudmunsson, 1986; Chin & Neu, 1987; Fuursted, 1987). However, Fuursted (1987) using sub-inhibitory concentrations of ciprofloxacin observed no suppression of growth i.e., no PAE, after drug removal with P. aeruginosa. In contrast the data from this study shows that at 0-5 x MIC of ciprofloxacin and enoxacin has a PAE of 0-9 h and 0-7 h on E. coli J96. CI-960, which had the lowest MIC of the three quinolones used in this study, did not show a significant PAE at 0-5 x MIC. The data shows that the filamentation of E. coli can be induced by exposure of 0-5 x MIC of CI-960, ciprofloxacin and enoxacin. The mechanism of cell division inhibition and subsequent filamentation of quinolone treated cells is due to induction of the SOS response, an error-prone DNA repair pathway (Dougherty & Saukkonen, 1985; Diver & Wise, 1986; Phillips et al., 1987; Piddock & Wise, 1987). The results obtained in this study suggest that the SOS response did not stop immediately after drug removal, as the quinolone treated cells continued to form filaments and double their length over a
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
X elongated from 30 to 65 /an during this time period. At time 120 min, the original cell divided and formed two daughter cells: XI and X2. The times for secondary cell division of the two daughter cells were very different. XI took about 15 min to form its new daughter cells: X1A and X1B. There was no cell division observed on cell X2 even after 90 min (time 210 min). There were three division sites observed on cell X1A at time 150 min. After 60 min, at time 210 min, cell septation was only completed at one of these three division points. A similar phenomenon was also observed on cell XIB. Cell division was observed at 135 min, but the two divided cells could not separate after a further 75 min, i.e., at time 210 min. Under the same conditions, the first cell division of the control population only took 30 min. The associated cell septation was completed in an additional 15 min. The new daughter cells grew evenly and showed a mean cell division time of about 45 min. Visual cell septation in control cells never took longer than 30 min after division.
536
L. Goan and J. C. Barnham
Figure 2. The effect! of 0-5 x MIC of enoxarin and co-trimoxazole on external ((a), (b)) and internal ((c), (d)) haemotysin activity in E. eoli J96 during 1-h treatment, and after drug removal, (a), (c) O, control; • , enoxacin; (b), (d) O, control; • , co-trimoxazole.
period three times the time of normal cell division. The first cell division of the original elongated cells is often asymmetrical. The effect of quinolone on bacterial morphology lasted considerably longer than growth suppression, i.e., at least 4 h for some E. coli cells. Sub-inhibitory concentrations of the quinolones did not show bactericidal activity, however, at these concentrations phenotypic effects (PAE and bacterial filamentation) were observed, which may be related to bacterial DNA gyrase inhibition. Recently, Aleixandre et al. (1991) reported that low concentrations of riprofloxacin could effect the supercoiling activity of DNA gyrase on plasmid DNA. In-vivo and in-vitro studies have shown that expression of some genes is sensitive to inhibitors of DNA gyrase (Shuman & Schwartz, 1975; Kano et al., 1981). Quinolones affect the functions of
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
8
PAEofqntootoots
537
Acknowledgements We thank the Parke-Davis Pharmaceutical Research Division, Warner-Lambert Company for their research support. References Aleixandre, V., Herrera, G., Urios, A. & Blanco, M. (1991). Effects ofriprofloxacinon plasmid DNA supercoiling of Escherichia coll topoisomerase I and gyrase mutants. Antimicrobial Agents and Chemotherapy 35, 20-3. Burnham, J. C. & Sonstein, S. A. (1990). Anti-virulence effects of lomefloxacin and other quinolones. In Program and Abstracts of the Thirtieth Interscienct Conference on Antimicrobial Agents and Chemotherapy, Atlanta, GA, 1990. Abstract 1010, p. 254. American Society for Microbiology, Washington, DC. Cavalicri, S. J., Bohach, G. A. & Snyder, I. S. (1984). Escherichia coll ot-hemolysin: characteristics and probable role in pathogenicity. Microbiological Reviews, 48, 326-43. Chin, N. X. & Neu, H. C. (1987). Post-antibiotic suppressive effect of ciprofloxacin against Gram-positive and Gram-negative bacteria. American Journal of Medicine 82, Suppl. 4A, 58-62. Craig, W. A. & Gudmunsson, S. (1986). The postantibiotic effect In Antibiotics in Laboratory Medicine 2nd ed. (Lorian, V., Ed.), pp. 515-36. Williams & Willrins, Baltimore, MD. Dalhoff, A. & Ddring, G. (1985). Interference of ciprofloxacin with the expression of pathogenicity factors of Pseudomonas aerugtnosa. In 77K Influence of Antibiotics on the
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
gyrase, and may alter the DNA structure, and further effect bacterial gene expression (Yang et al., 1979). The expression of the haemolysin gene, a virulence associated factor, appears to be one of the genes affected by quinolones. The effects of subinhibitory concentrations of quinolones on other extra-cellular factors (e.g., exotoxin; Dalhoff & Doring, 1985) or structure factors (e.g., fimbriae; Burnham & Sonstcin, 1990; Guan, Drill & Burnham, 1991) may be based on the same mechanism. Previous studies have demonstrated that exposure to low concentrations of quinolones may reduce the ability of pathogens to cause clinical symptoms via a reduction in the level of production of virulence factors (Burnham & Sonstein, 1990). The role of haemolysin in the virulence of E. coli has been demonstrated in various animal models and cell culture (Cavalier et al., 1984). Both external and internal haemolysin activity was examined in this study in order to determine whether transport of the enzyme might be defective during PAE period. The external haemolysin activity was not significantly affected after 1 h exposure to enoxacin. This could be because the exposure time was not long enough to cause an affect, or that there was substantial external haemolysin secreted in the medium and internal haemolysin inside E. coli cells before drug exposure. Following quinolone removal the treated cells were unable to exhibit normal haemolysin activity for at least 2 h. Internal haemolysin activity was only adversely affected for 1 h. This suggests that transport of the enzyme from the cytoplasm to the external environment was probably unaffected since it would take approximately the extra hour for the concentration of the haemolysin to accumulate sufficiently for the assay to detect it extracellularly. The results of this study suggest that the consideration of postantibiotic effects should include the residual antibiotic effects on bacterial morphology and virulence factors, in addition to the suppression of bacterial regrowth. We believe the excellent in-vivo activity of quinolone antibiotics could be partly due to their extended PAE effects on exposed bacterial cells.
538
L- Gran and J. C. Barnham
(Received 24 July 1991; revised version accepted 20 December 1991)
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Host-Parasite Relationship II (Adam, D., Hahn, H. & Opferkuch, W., Eds), pp. 246-55. Springer-Verlag, Berlin. Diver, J. M. & Wise, R. (1986). Morphological and biochemical changes in Escherichia coli after exposure to ciprofloxacin. Journal of Antimicrobial Chemotherapy 18, Suppl. D, 31-41. Dougherty, T. J. & Saukkonen, J. J. (1985). Membrane permeability changes associated with DNA gyrase inhibitors of Escherichia coli. Antimicrobial Agents and Chemotherapy 28, 200-6. Eliopoulos, G. M. & Eliopoulos, C. T. (1989). Quinolone antimicrobial agents: activity in vitro. In Quinolone Antimicrobial Agents (Wolfson, J. S. & Hooper, D. C , Eds), pp. 35-70. American Society for Microbiology, Washington, DC. Fuursted, K. (1987). Post-antibiotic effect of ciprofloxacin on Pseudomonas aeruginosa. European Journal of Clinical Microbiology 6, 271-4. Guan, L., Drill, C. L. & Burnham, J. C. (1991). Antiadherence and postantibiotic effects (PAE) of CI-960 and enoxacin on Escherichia coli. Program and Abstracts of the Thirty-First Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, 1991. Abstract 1135, p. 288. American Society for Microbiology, Washington, DC. Hays, J. B. & Boehmer, S. (1978). Antagonists of DNA gyrase inhibit repair and recombination of UV-irradiated phage lambda. Proceedings of the National Academy of Sciences of the USA 75, 4125-9. Kano, Y., Miyashita, T , Nakamura, H., Kuroki, K., Nagata, A. & Imamoto, F. (1981). In vivo correlation between DNA supercoiling and transcription. Gene 13, 173-84. Kreuzer, K. N. & Cozzarelli, N. R. (1979). Escherichia coli mutants thermosensitive for deoxyribonucleic acid gyrase subunit A: effects on deoxyribonucleic acid replication, transcription, and bacteriophage growth. Journal of Bacteriology 140, 424-35. Normark, S., Lark, D., Hull, R., Norgren, M., BAga, M., O'Hantey, P. et al. (1983). Genetics of digalactoside-binding adhesin from a uropathogenic Escherichia coli strain. Infection and Immunity 41, 942-9. Ostrowski, J. & Hulanicka, D. (1981). Effect of DNA gyrase inhibitors on gene expression of the cysteine regulon. Molecular and General Genetics 181, 363-6. Phillips, I., Culebras, E., Moreno, F. & Baquero, F. (1987). Induction of the SOS response by new 4-quinolones. Journal of Antimicrobial Chemotherapy 20, 631-8. Piddock, L. J. V. & Wise, R. (1987). Induction of the SOS Response in Escherichia coli by 4-quinolone antimicrobial agents. FEMS Microbiology Letters 41, 289-94. Sanzey, B. (1979). Modulation of gene expression by drugs affecting deoxyribonucleic acid gyrase. Journal of Bacteriology 138, 40-7. Shuman, H. & Schwartz, M. (1975). The effect of nalidixic acid on the expression of some genes in Escherichia coli K-12. Biochemical and Biophysical Research Communications 64, 204-9. Springer, W. & Goebel, W. (1980). Synthesis and secretion of hemolysin by Escherichia coli. Journal of Bacteriology 144, 53-9. Vogehnan, B., Gudmundsson, S., Turnidge, J., Leggett, J. & Craig, W. A. (1988). In vivo postantibiotic effect in a thigh infection in neutropenic mice. Journal of Infectious Diseases 157, 287-98. Wheat, R. W. (1988). Bacterial morphology and ultrastructure. In Zinsser Microbiology, 19th edn (Joklik, W. K., Willett, H. P., Amos, D. B. & Wilfcrt, C. M., Eds), p. 15. Appleton & Lange, East Norwalk, CT. Yang, H. L., Heller, K., Gellert, M. & Zubay, G. (1979). Differential sensitivity of gene expression in vitro to inhibitors of DNA gyrase. Proceedings of the National Academy of Sciences of the USA 76, 3304-8.
Postantibiotic effect of CI-960, enoxacin and dprofloxacin on Escherichia coir, effect on morphology and haemolysin activity
Department of Microbiology, Medical College of Ohio, Toledo, Ohio 43699, USA The postantibiotic effect (PAE) has been classically defined as the suppression of bacterial growth that persists after limited exposure of organisms to antimicrobial agents. Morphology and haemolysin activity during the PAE of three quinolones on Escherichia coli were examined in this study. A one hour exposure to the quinolones, CI-960, enoxacin and dprofloxacin, produced a PAE of 0-5-2-0 h. When determinated by Coulter counter, at 0-5 x MIC of enoxacin or CI-960 after 1 h exposure, 58% or 42% cells, respectively, of the treated cells were filamentous (cell length > 12 fan). After drug removal, the population of the filamentous cells decreased, however, after even 4 h, 12% and 2% of the cells were still filamentous after exposure to enoxacin or CI-960. Further morphological studies during the PAE showed that the first division of the filamentous cell was asymmetrical, and both bacterial cell division and septation were delayed after exposure to 0-5 MIC of CI-960. Following quinolone removal, the treated E. coli did not exhibit normal activity of haemolysin for at least 2 h. Internal haemolysin activity was adversely affected for 1 h. The results of this study suggest that any consideration of postantibiotic effects should include the residual antibiotic effects on bacterial morphology and virulence factors, in addition to the defined suppression of bacterial regrowth.
Introduction The postantibiotic effect (PAE) has been classically defined as a term used to describe the suppression of bacterial growth that persists after limited exposure of organisms to antimicrobial agents (Vogelman et al., 1988). The longer PAE, the lower the probability that growth of the infecting bacterium will resume during that fraction of the interdose interval where antimicrobial levels fall below the minimum inhibitory concentration (MIC) (Eliopoulos & Eliopoulos, 1989). A long PAE provides the potential for administering the antimicrobial agent at longer intervals between doses (Craig & Gudmunsson, 1986). This classically defined PAE usually only addresses the bacterial regrowth after a short exposure to an antimicrobial agent. Other bacterial physiological characters (especially virulence factors) during the PAE have not been well studied. If bacterial virulence factor production or activity decreased after exposure to antibiotics, the clinical relevance of PAE could be greater than previously considered. The primary target site for quinolones is DNA gyrase, which has been shown to be involved in a number of cellular processes other than supercoiling of bacterial DNA. DNA replication, transcription, and repair as well as recombination have been shown to be DNA-gyrase dependent (Hays & Bocchmer, 1978; Kreuzer & Cozzarclli, 1979). 529 0305-7453/92/050529 +10 $02.00/0
© 1992 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Iinong Gun and Jeffrey C. Burnham
530
L. Gnu and J. C Burnham
Materials and methods
Bacterial strains E. coli strains ATCC 25922 and J96 (O4, K6; provided by Sheila Hull, Baylor College of Medicine, Houston, TX), was isolated from a patient with acute pyelonephritis (Normark et al., 1983). Determination of MIC and MBC The MIC of CI-960 and enoxacin (Parke Davis Pharmaceutical Research Division, Warner-Lambert Company, Ann Arbor, MI), and ciprofloxacin and co-trimoxazole (Sigma Chemical Co., St. Louis, MO) for each strain were determined by microbroth dilution in Mueller-Hinton Broth (Difco). After determination of the approximate MIC, a more accurate MIC was determined by dividing the range between the 0-5 x MIC and 2 x MIC into ten dilutions. The MIC determinations were read after incubation at 37°C for 18-20 h. An inoculum of approximately 1 x 103 cfu/mL was used. The bacterial inoculum for MBC determination was prepared as above for the determination of the MICs. After 18-20 h of incubation with the drugs and 10- to 100fold dilution in sterile 0-85% N a d , bacteria were counted by subculturing 0-1 mL of bacteria directly onto Mueller-Hinton Agar (Difco) for evaluation after 18-20 h at 37°C. Experiments were performed in duplicate. The MBC was taken as the lowest concentration which gave 99-9% bacterial killing. PAE The reference tube macrodilution method was performed as described by Craig & Gudmunsson (1986). The inocula were prepared by incubating the bacteria overnight at 37°C in Mueller-Hinton broth. The culture was then diluted 1:100 in fresh prewarmed Mueller-Hinton broth and incubated in a shaking water bath until the
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Evidence has accumulated indicating that at higher doses, the quinolones and/or novobiocin interfere with transcription-translation of genes sensitive to catabolite repression but have no effect on constitutively expressed genes (Sanzey, 1979; Ostrowslri & Hulanicka, 1981; Dalhoff & Doring, 1985). The synthesis of virulence factors, e.g. proteases and exotoxin A in Pseudomonas aeruginosa was significantly reduced by subinhibitory concentrations of dprofloxatin under in-vitro as well as invivo conditions. The degree of inhibition of protease and toxin production was strainspecific but independent of metabolic inhibition (Dalhoff & Doring, 1985). Similar effects by quinolones at subinhibitory concentrations on virulence-associated factors in Escherichia coli have also been reported (Burnham & Sonstein, 1990). Chin & Neu (1987) showed that ciprofloxacin had a 3- to 6-h postantibiotic suppressive effect on growth of staphylococci, Enterobacteriaceae and P. aeruginosa, but not on Enterococcus faecalis. The aim of this work was to investigate and compare the PAEs of CI-960 (PD127391), a newer quinolone, along with two well studied quinolones, ciprofloxacin and enoxacin, with that of co-trimoxazole on E. coli. The response of bacterial morphology and haemolysin activity during the PAE period was also studied.
PAE of qninolones
531
Morphology The effects of antibiotics on bacterial morphology were determined by using a Coulter counter (Model ZM, Technical Communications, Coulter Electronics, Inc., USA). The population of the cells with abnormal size (cell length > 12 /an) were measured after exposure to antibiotics and subsequent drug removal. Bacterial cell division and septation were studied after exposure to 0-5 x MIC CI-960 for 1 h. The drug was removed by washing with fresh Mueller-Hinton broth and the washed cells were incubated on the surface of Mueller-Hinton agar. Bacterial morphological changes were monitored every 15 min by phase contrast microscopy (Axiophot, Carl Zeiss Instruments, Inc., Germany). Haemolysin activity Haemolysin activity was measured as described by Springer & Goebel (1980) with the modifications described as follows: Accumulation of haemolysin: 1 mL samples of E. coli cells were taken at intervals as described for the measurement of PAE. The cells were removed by centrifugation, and the supernatants assayed for haemolytic activity which was denned as 'external haemolysin'. The washed cell pellets were suspended in 1 mL of 0-01 M phosphate buffer (pH 6-0) and sonicated. The cell debris was removed . by centrifugation, and the supernatants were tested for haemolytic activity and defined as 'internal haemolysin'. Haemolytic activity was determined by the amount of haemoglobin released from washed erythrocytes: one hundred microlitres of external or internal haemolysin was added to 100 pL ox erythrocytes (buffered by pH 7-5, 0-01 M Tris-hydrochloride, 002 M calcium chloride) in microtitre plates. The assay plates were incubated at 40°C for 10 min, and incubation was terminated by quickly removing unlysed erythrocytes by centrifugation at 4°C. The amount of released haemoglobin was measured by the absorbance of the supernatants at 405 nm. One unit of haemolysin activity was defined as the amount of haemolysin protein which caused the release of 1 /anol of haemoglobin in 1 h under test conditions. Results MICs and MBCs The MICs and MBCs of each antibiotic for E. coli ATCC 25922 and J96 are given in Table I. CI-960 was the most active with the lowest MIC and MBC; the MBC being double the MIC. The MICs of ciprofloxacin and enoxacin were equal to the MBCs.
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
logarithmic phase of growth was reached. Cultures with a cell density of approximately 5x10* cfu/mL were then exposed to various drug concentrations for 1 h. In order to remove residual antibiotics completely, cultures were diluted 100- or 1000-fold in fresh pre-warmed Mueller-Hinton broth. Viable counts were determined on Mueller-Hinton agar plates for the initial inoculum, following the removal of the antibiotics. The PAE was calculated from the equation PAE (time) •= T-C, where T is the time required for the count in cfu/mL of the test culture to increase by one log10 above the count observed immediately after drug removal, and C is the time required for the count of cfu/mL in an untreated control culture to increase one log,0 above the count observed immediately after completion of the same procedure used on the culture for drug removal (Craig & Gudmunsson, 1986).
L. Guan and J. C Burnham
532
Table L Susceptibility to and PAE of E. coli exposed for 1 h to four antibiotics in different concentrations
MIC Strain ATCC 25922
(mg/L)
0004 0015
CI-960 ciprofloxacin enoxacin co-trimoxazole*
0*5
Q-960
O004
ciprofloxacin
003
enoxacin
019
co-trimoxazole
O60
0-16
Concentration (mg/L)
0016 (4 x MIQ 006 (4xMIQ 064 (4 x MIQ 2-40 (4 x MIQ 0016 (4 x MIQ O004 (1 x MIQ 0002 (05 x MIQ O001 (025 x MIQ 012 (4 x MIQ 003 (1 x MIQ 0015 (05 x MIQ O008 (025 x MIQ 076 (4 x MIQ 019 ( l x MIQ O095(O5xMIQ 0048 (025 x MIQ 2-40 (4 x MIQ O60 (1 x MIQ
030 (05 x MIQ
014 (025 x MIQ
PAE (h) 1-7 1-6
09 02 20 05 01 -01 1-8 1-2 09 05 1-4 1-2 07 02 02 00 00 ND
'1 trimethoprim: 5 lulfamethoxazole. ND, Not done.
PAE The PAE of the four drugs in different concentrations for the two strains of E. coli after exposure for 1 h is shown in Table I. A significant period of growth suppression was seen after exposure of E. coli ATCC 25922 and J96 to CI-960, ciprofloxacin and enoxacin at a concentration of 4 x MIC. For E. coli J96, CI-960 showed the longest PAE, 2-0 h, at 4 x MIC, and shortest PAE, 0-5 h, at 1 x MIC; whereas ciprofloxacin and enoxacin produced a PAE of 0-9 h and 0-7 h respectively, at 0-5 x MIC. CI-960 did not show significant PAE at sub-inhibitory concentrations. PAE was not seen with E. coli ATCC 25922 at sub-inhibitory concentrations with any of the quinolones. Co-trimoxazole did not show significant PAE at concentrations up to 4 x MIC with either strains. CI-960, ciprofloxacin and enoxacin were bactericidal at 4 x MIC for E. coli J96 after 1 h exposure (Table II). Enoxacin and ciprofloxacin also showed bacterial killing activity at 1 x MIC. All the drugs tested showed no bactericidal activity at concentrations below their MICs (Table II).
Morphology In order to avoid the interference from dead cells, quinolones were used at 0-5 x MIC for bacterial morphology and haemolysin activity assays.
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
J96
Antibiotic
533
PAE of qatnotonei
Table IL The viable count of E. coli J96 after one hour's exposure to CI-960, enoxadn and ciprofloxacin Antibiotic
Concentration (mg/L)
Viable cell count before exposure after exposure 8x10*
None CI-960
Ciprofloxacin
8xl07 2x10* 2x10'
99-75 0
8x10' 8x10' 2x10* 1x10* 2x10' 8x10'3
0
3X1O
0 97-50 87-00 0
0
4x10* 2x10'
99-96 50-00 0
8x10'
0
These morphological studies resulted in two major findings: (i) While the normal E. coli cell length is 1-3 fan (Wheat, 1988), filamentation of E. coli cells is a common phenomenon when exposed to quinolones (Diver & Wise, 1986). In this study the morphological changes of these filamentous cells after the antibiotic was removed were examined. Table i n shows the effects of 0-5 x MIC of CI-960 and enoxacin on E. coli J96 morphology as determined by Coulter counter. Only 0-12% of the population of antibiotic-free cells are longer than 12 fan. At 0-5 x MIC of enoxacin or CI-960, after exposure 1 h, 58% and 42% cells, respectively, were longer than 12 fan. After drug removal, the population of filamentous cells decreased. Even after 4 h, 12% of the enoxacin exposed cells were still filamentous, and 2% filamented of those treated with CI-960. These results show that the effects of CI-960 and enoxacin on E. coli morphology can last at least 4 h. Ciprofloxacin showed similar effects to enoxacin. Co-trimoxazole did not show significant effects at 0-5 x MIC. (ii) Figure 1 shows bacterial cell division and septation after exposure to 0-5 x MIC of CI-960. Cell 'X' was one of the treated cells. At time 0, just after washing, the length of cell X was 30 fan. There was no cell division observed after 105 min, although cell TaMe m. Effects of 0-5 x the MIC of CI-960 or enoxacin on E. coli J96 morphology (measured by Coulter counter) Time (h) 0* 1 2
3 4
Control CI-960 Enoxacin cells total cells cells total cells total cells cells counted > 12 urn (%) counted > 12 urn (%) counted > 12 um (%) 1-4X104 2-OxlO4 1-6X104 1-7 xlO 4 1-3 xlO 4
12 (0-12) 5(0-025) 0(0) 0(0)
0(0)
4-0 xlO 4 1-2 xlO 4 1-7 xlO 4 4-2 x 104
51 x 10*.
16810 (42)
6000(50) 3061 (18)
3362 (8) 102(2)
5-3 xlO 3 7-6 xlO 3 4-2 xlO 4 9-2 xlO 4 4-4x10*
3074 (58) 3648 (42) 15690 (38) 1840(20) 5280 (12)
Time 0 - after 1 b exposure to 0-5 x MIC of 0-960 or enoxacin and subsequent drug removal.
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Enoxacin
(K)16(4xMIQ 0O04 (1 x MIC) 0-002 (0-5 x MIC) 0O01 (0-25 x MIC) 0-76 (4 x MIC) 0-19 l x M I Q (H)95 (0-5 x MIC) (MM8 (0-25 x MIC) 0-12 (4 x MIC) 003 (1 x MIC) 0-015 (0-5 x MIC) 0-008 (0-25 x MIC)
Killing (%)
I
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
(c)
Figure 1. Light micrograph of E. coli J96 cell division and septation after exposure to 0-5 x the MIC of CI-960: (a) cell X at 0 min, PAE, i.e., after 1 h exposure and subsequent washing with Mueller-Hinton broth to remove the drug; (b) at 105 min; (c) at 120 min, cell XI and X2 are daughter cells of cell X; (d) at 135 min, cell XIA and XIB are daughter cells of cell XI; (e) at 150 min; (f) at 210 min, cell XlAl and X1A2 are daughter cells of cell XIA. Cell length: 1 cm = 7-5 fan.
PAE of qriookraes
535
Haemolysin To compare the haemolysin activity between the antibiotic-free control and antibiotictreated cells, cell mass (e.g., 'per O D ^ ' (unit of optical density at 520 nm)) was used as a reference unit Figure 2 shows the effects of 0-5 x MIC of enoxacin and cotrimoxazole on both external and internal haemolysin activity in E. coli J96 after removal of drug. The external haemolysin activity of 0-5 x MIC of enoxacin treated cells was inhibited for at least 2 h, about 1 h longer than PAE (the PAE was 0-7 h, Table I). One hour and 2 h after drug removal, the external haemolysin activity of the treated cells was about 56% and 65% of the activity of control cells (Figure 2(a)). The internal haemolysin activity of treated cells was strongly inhibited for 1 h. The activity of internal haemolysin of the treated cells was about 31 % of the control cells after drug removal at 1 h (Figure 2(c)). Ciprofloxacin showed similar effects to enoxacin on both external and internal haemolysin activity. Co-trimoxazole did not show any significant effect on the haemolysin activity of E. coli (Figure 2(b), (d)).
Discussion This study shows that CI-960, ciprofloxacin and enoxacin can produce significant PAE on E. coli. An increase in the concentration of the three quinolones was associated with prolongation of PAE. This is similar to data reported by other workers (Craig & Gudmunsson, 1986; Chin & Neu, 1987; Fuursted, 1987). However, Fuursted (1987) using sub-inhibitory concentrations of ciprofloxacin observed no suppression of growth i.e., no PAE, after drug removal with P. aeruginosa. In contrast the data from this study shows that at 0-5 x MIC of ciprofloxacin and enoxacin has a PAE of 0-9 h and 0-7 h on E. coli J96. CI-960, which had the lowest MIC of the three quinolones used in this study, did not show a significant PAE at 0-5 x MIC. The data shows that the filamentation of E. coli can be induced by exposure of 0-5 x MIC of CI-960, ciprofloxacin and enoxacin. The mechanism of cell division inhibition and subsequent filamentation of quinolone treated cells is due to induction of the SOS response, an error-prone DNA repair pathway (Dougherty & Saukkonen, 1985; Diver & Wise, 1986; Phillips et al., 1987; Piddock & Wise, 1987). The results obtained in this study suggest that the SOS response did not stop immediately after drug removal, as the quinolone treated cells continued to form filaments and double their length over a
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
X elongated from 30 to 65 /an during this time period. At time 120 min, the original cell divided and formed two daughter cells: XI and X2. The times for secondary cell division of the two daughter cells were very different. XI took about 15 min to form its new daughter cells: X1A and X1B. There was no cell division observed on cell X2 even after 90 min (time 210 min). There were three division sites observed on cell X1A at time 150 min. After 60 min, at time 210 min, cell septation was only completed at one of these three division points. A similar phenomenon was also observed on cell XIB. Cell division was observed at 135 min, but the two divided cells could not separate after a further 75 min, i.e., at time 210 min. Under the same conditions, the first cell division of the control population only took 30 min. The associated cell septation was completed in an additional 15 min. The new daughter cells grew evenly and showed a mean cell division time of about 45 min. Visual cell septation in control cells never took longer than 30 min after division.
536
L. Goan and J. C. Barnham
Figure 2. The effect! of 0-5 x MIC of enoxarin and co-trimoxazole on external ((a), (b)) and internal ((c), (d)) haemotysin activity in E. eoli J96 during 1-h treatment, and after drug removal, (a), (c) O, control; • , enoxacin; (b), (d) O, control; • , co-trimoxazole.
period three times the time of normal cell division. The first cell division of the original elongated cells is often asymmetrical. The effect of quinolone on bacterial morphology lasted considerably longer than growth suppression, i.e., at least 4 h for some E. coli cells. Sub-inhibitory concentrations of the quinolones did not show bactericidal activity, however, at these concentrations phenotypic effects (PAE and bacterial filamentation) were observed, which may be related to bacterial DNA gyrase inhibition. Recently, Aleixandre et al. (1991) reported that low concentrations of riprofloxacin could effect the supercoiling activity of DNA gyrase on plasmid DNA. In-vivo and in-vitro studies have shown that expression of some genes is sensitive to inhibitors of DNA gyrase (Shuman & Schwartz, 1975; Kano et al., 1981). Quinolones affect the functions of
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
8
PAEofqntootoots
537
Acknowledgements We thank the Parke-Davis Pharmaceutical Research Division, Warner-Lambert Company for their research support. References Aleixandre, V., Herrera, G., Urios, A. & Blanco, M. (1991). Effects ofriprofloxacinon plasmid DNA supercoiling of Escherichia coll topoisomerase I and gyrase mutants. Antimicrobial Agents and Chemotherapy 35, 20-3. Burnham, J. C. & Sonstein, S. A. (1990). Anti-virulence effects of lomefloxacin and other quinolones. In Program and Abstracts of the Thirtieth Interscienct Conference on Antimicrobial Agents and Chemotherapy, Atlanta, GA, 1990. Abstract 1010, p. 254. American Society for Microbiology, Washington, DC. Cavalicri, S. J., Bohach, G. A. & Snyder, I. S. (1984). Escherichia coll ot-hemolysin: characteristics and probable role in pathogenicity. Microbiological Reviews, 48, 326-43. Chin, N. X. & Neu, H. C. (1987). Post-antibiotic suppressive effect of ciprofloxacin against Gram-positive and Gram-negative bacteria. American Journal of Medicine 82, Suppl. 4A, 58-62. Craig, W. A. & Gudmunsson, S. (1986). The postantibiotic effect In Antibiotics in Laboratory Medicine 2nd ed. (Lorian, V., Ed.), pp. 515-36. Williams & Willrins, Baltimore, MD. Dalhoff, A. & Ddring, G. (1985). Interference of ciprofloxacin with the expression of pathogenicity factors of Pseudomonas aerugtnosa. In 77K Influence of Antibiotics on the
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
gyrase, and may alter the DNA structure, and further effect bacterial gene expression (Yang et al., 1979). The expression of the haemolysin gene, a virulence associated factor, appears to be one of the genes affected by quinolones. The effects of subinhibitory concentrations of quinolones on other extra-cellular factors (e.g., exotoxin; Dalhoff & Doring, 1985) or structure factors (e.g., fimbriae; Burnham & Sonstcin, 1990; Guan, Drill & Burnham, 1991) may be based on the same mechanism. Previous studies have demonstrated that exposure to low concentrations of quinolones may reduce the ability of pathogens to cause clinical symptoms via a reduction in the level of production of virulence factors (Burnham & Sonstein, 1990). The role of haemolysin in the virulence of E. coli has been demonstrated in various animal models and cell culture (Cavalier et al., 1984). Both external and internal haemolysin activity was examined in this study in order to determine whether transport of the enzyme might be defective during PAE period. The external haemolysin activity was not significantly affected after 1 h exposure to enoxacin. This could be because the exposure time was not long enough to cause an affect, or that there was substantial external haemolysin secreted in the medium and internal haemolysin inside E. coli cells before drug exposure. Following quinolone removal the treated cells were unable to exhibit normal haemolysin activity for at least 2 h. Internal haemolysin activity was only adversely affected for 1 h. This suggests that transport of the enzyme from the cytoplasm to the external environment was probably unaffected since it would take approximately the extra hour for the concentration of the haemolysin to accumulate sufficiently for the assay to detect it extracellularly. The results of this study suggest that the consideration of postantibiotic effects should include the residual antibiotic effects on bacterial morphology and virulence factors, in addition to the suppression of bacterial regrowth. We believe the excellent in-vivo activity of quinolone antibiotics could be partly due to their extended PAE effects on exposed bacterial cells.
538
L- Gran and J. C. Barnham
(Received 24 July 1991; revised version accepted 20 December 1991)
Downloaded from http://jac.oxfordjournals.org/ at Thomas Jefferson University, Scott Library on March 9, 2015
Host-Parasite Relationship II (Adam, D., Hahn, H. & Opferkuch, W., Eds), pp. 246-55. Springer-Verlag, Berlin. Diver, J. M. & Wise, R. (1986). Morphological and biochemical changes in Escherichia coli after exposure to ciprofloxacin. Journal of Antimicrobial Chemotherapy 18, Suppl. D, 31-41. Dougherty, T. J. & Saukkonen, J. J. (1985). Membrane permeability changes associated with DNA gyrase inhibitors of Escherichia coli. Antimicrobial Agents and Chemotherapy 28, 200-6. Eliopoulos, G. M. & Eliopoulos, C. T. (1989). Quinolone antimicrobial agents: activity in vitro. In Quinolone Antimicrobial Agents (Wolfson, J. S. & Hooper, D. C , Eds), pp. 35-70. American Society for Microbiology, Washington, DC. Fuursted, K. (1987). Post-antibiotic effect of ciprofloxacin on Pseudomonas aeruginosa. European Journal of Clinical Microbiology 6, 271-4. Guan, L., Drill, C. L. & Burnham, J. C. (1991). Antiadherence and postantibiotic effects (PAE) of CI-960 and enoxacin on Escherichia coli. Program and Abstracts of the Thirty-First Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, IL, 1991. Abstract 1135, p. 288. American Society for Microbiology, Washington, DC. Hays, J. B. & Boehmer, S. (1978). Antagonists of DNA gyrase inhibit repair and recombination of UV-irradiated phage lambda. Proceedings of the National Academy of Sciences of the USA 75, 4125-9. Kano, Y., Miyashita, T , Nakamura, H., Kuroki, K., Nagata, A. & Imamoto, F. (1981). In vivo correlation between DNA supercoiling and transcription. Gene 13, 173-84. Kreuzer, K. N. & Cozzarelli, N. R. (1979). Escherichia coli mutants thermosensitive for deoxyribonucleic acid gyrase subunit A: effects on deoxyribonucleic acid replication, transcription, and bacteriophage growth. Journal of Bacteriology 140, 424-35. Normark, S., Lark, D., Hull, R., Norgren, M., BAga, M., O'Hantey, P. et al. (1983). Genetics of digalactoside-binding adhesin from a uropathogenic Escherichia coli strain. Infection and Immunity 41, 942-9. Ostrowski, J. & Hulanicka, D. (1981). Effect of DNA gyrase inhibitors on gene expression of the cysteine regulon. Molecular and General Genetics 181, 363-6. Phillips, I., Culebras, E., Moreno, F. & Baquero, F. (1987). Induction of the SOS response by new 4-quinolones. Journal of Antimicrobial Chemotherapy 20, 631-8. Piddock, L. J. V. & Wise, R. (1987). Induction of the SOS Response in Escherichia coli by 4-quinolone antimicrobial agents. FEMS Microbiology Letters 41, 289-94. Sanzey, B. (1979). Modulation of gene expression by drugs affecting deoxyribonucleic acid gyrase. Journal of Bacteriology 138, 40-7. Shuman, H. & Schwartz, M. (1975). The effect of nalidixic acid on the expression of some genes in Escherichia coli K-12. Biochemical and Biophysical Research Communications 64, 204-9. Springer, W. & Goebel, W. (1980). Synthesis and secretion of hemolysin by Escherichia coli. Journal of Bacteriology 144, 53-9. Vogehnan, B., Gudmundsson, S., Turnidge, J., Leggett, J. & Craig, W. A. (1988). In vivo postantibiotic effect in a thigh infection in neutropenic mice. Journal of Infectious Diseases 157, 287-98. Wheat, R. W. (1988). Bacterial morphology and ultrastructure. In Zinsser Microbiology, 19th edn (Joklik, W. K., Willett, H. P., Amos, D. B. & Wilfcrt, C. M., Eds), p. 15. Appleton & Lange, East Norwalk, CT. Yang, H. L., Heller, K., Gellert, M. & Zubay, G. (1979). Differential sensitivity of gene expression in vitro to inhibitors of DNA gyrase. Proceedings of the National Academy of Sciences of the USA 76, 3304-8.
