Journal of Antimicrobial Chemotherapy (1990) 25, 551-560
Synergy between penicillin and gentamidn against enterococci T. G. Wlnstanky* and J. G. M. Hastings6*
The role of active uptake in aminoglycoside activity against penicillin-treated enterococci was studied by viable counts and ATP determinations. Penicillin and gentamicin gave syncrgistic bactericidal and post-antibiotic effects (PAEs) which were partially reduced by sodium azide, an electron transport inhibitor, and totally blocked in the presence of both sodium azide and EDTA, which chelates divalent cations. EDTA and gentamicin showed marked synergy in both 'killing curve' and PAE experiments. This synergy was completely inhibited by sodium azide. The data indicate that the activity of gentamicin against enterococci that have been damaged by penicillin or EDTA is energy-dependent This is consistent with present theories of gentamicin uptake via transportation driven by a protonmotive force.
Introduction
Synergy between penicillins and aminoglycosides was first reported by Hunter (1947), and the combination of penicillin and streptomycin has been a classic example of antibiotic synergy since it was first described in relation to Escherichia coli (Anand, Davis & Armitage, 1961). Plotz & Davis (1962) demonstrated that penicillin stimulated the uptake of streptomycin by Esch. coli. Other cell wall active agents, such as vancomycin, increase the uptake of aminoglycosides by enterococci (Moellering & Weinberg, 1971). The uptake and lethal effects of streptomycin require active respiration (Hancock, 1962) and, in Esch. coli, two energy-dependent phases of streptomycin uptake, both inhibited by respiratory poisons, have been described (Bryan & Van den Elzen, 1977). The role of active transport in aminoglycoside uptake by antibiotic-damaged Gram positive organisms is less clear. Yee, Farber & Mates (1986) showed that sodium azide, an inhibitor of electron transport, could prevent the increase in streptomycin uptake seen in viridans streptococci that had been pretreated with cell wall synthesis inhibitors. However, recent work (Fuursted, 1989) has contradicted these observations, indicating that energy is not required for aminoglycoside uptake by enterococci that have been pretreated with cell wall synthesis inhibitors. The aim of the present work was to resolve these conflicting findings by studying the interaction of penicillin and gentamicin alone and in the presence of a chelating agent and a respiratory poison. •Present address: Department of Clinical Microbiology, The Queen Elizabeth Hospital, Queen Elizabeth Medical Centre, Edgbaston, Birmingham B15 2TH, UK. 551 03O5-7453/9O/WO551 + 10 S02.00/0
© 1990 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
'Department of Bacteriology, The Royal Hallamshire Hospital, Glossop Road, Sheffield, S102JF; bDepartment of Medical Microbiology, The University of Sheffield Medical School, Sheffield S102RX, UK
552
T. G. Winstanky and J. G. M. Hastings Materials and methods
Bacteria Enterococcus faecalis strain 281 (Winstanley & Hastings, 1989) was used throughout. The MICs of penicillin and gentamicin for this strain are 1 and 8 mg/1, respectively. The organism was stored in glyccrol broth at — 70°C and was subcultured on to blood agar when required. Antibacterial agents
Determination of bactericidal activity and post-antibiotic effect (PAE) Organisms were grown in Iso-Sensitest Broth supplemented with glucose (0-1% w/v) (ISB, Oxoid) to a concentration of 10*cfu/ml. This density was attained in midlogarithmic phase. Antibacterial agents were then added to the cultures to achieve the following concentrations: benzylpenicillin (10 mg/1), gentamicin (10 mg/1), EDTA (1500 mg/1) and sodium azide (BDH; 0-25% w/v). Untreated cultures were used as controls. Bactericidal activity was assessed by surface viable counts on blood agar (Miles, Misra & Irwin, 1938) immediately after addition of the various agents to organisms and after incubation for 4h at 37°C. Post-antibiotic effects (PAEs) were determined by removal of the antibacterial agents from the bacterial cells after incubation for 1 h. Cultures were centrifuged (3000 g for lOmin), washed three times, and resuspended in fresh pre-warmed broth. Controls not exposed to antimicrobial agents were treated in the same manner. Bacterial regrowth was quantitated by measurement of total bacterial ATP, extracellular ATP or from viable counts. In further experiments, PAEs were determined after exposure of organisms in a sequential manner. The cells were exposed to penicillin (10 mg/1) for 1 h, followed by centrifugation and washing, and then exposed to gentamicin (10 mg/1), sodium azide (0-25%) or EDTA (1500 mg/1) or to combinations of the three for a further 1 h before a final washing procedure and measurement of bacterial regrowth. Assay of bacterial ATP Total bacterial ATP was determined by a bioluminescence method and taken as a measure of bacterial biomass. A 200 /d sample of culture was mixed with an equal volume of extraction reagent, trichloroacetic acid (2-5%) containing 4DIM EDTA. After 2 min, 20 /xl of this extract was added to 200 /J of buffer (Tris acetate, EDTA) and 20 /J of ATP-monitoring reagent (luciferin-luciferase; Pharmacia) in the well of a white plastic microtitration tray (Amersham International, Little Chalfont, UK). Light emission was measured in a himinomcter (Amerlite Analyser). A 10-fA volume of ATP standard (0-01 mM, Pharmacia) was then added to each well and the light output was again measured. The ATP concentration in the samples was then calculated using the formula: P = IOO(R-B)/(S-R) where P was the ATP concentration in picomoles, R the test reading, S the reading after the addition of the ATP standard and B the
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Benzylpenicillin (Glaxo) and gentamicin (Roussel Laboratories) were obtained as standard pharmaceutical preparations. Fresh solutions were prepared in distilled water before each experiment.
Penkfllin-gentamldii synergy
553
background reading of the luminometer. Extracellular ATP was determined by assay of 10/J of culture without extraction; other additions being as described above. Quantitation of PAE PAE was calculated from the equation PAE = T— C where T was the time required for the ATP content of the test culture to increase by 5 picomoles or for the viable count to increase ten-fold above the value pertaining immediately after removal of antibacterial agent and C was the time required for the control culture to increase by a similar amount. Preliminary data showed this value to occur at a stage when the bacteria were in the logarithmic growth phase.
Results were expressed as the mean and standard deviation of three readings. Student's /-test was used to compare differences in bactericidal activity. Results Bactericidal effects
Strain 281 was incubated for 4 h in the presence of various combinations of penicillin, gentamicin, EDTA and sodium azide (Figure 1). EDTA alone had a bacteriostatic effect, penicillin was weakly bactericidal but gentamicin had little effect on bacterial growth. There was a moderate suppression of growth in the presence of sodium azide. The addition of sodium azide had no significant effect on the bactericidal activity of penicillin. As expected, the penicillin plus gentamicin combination was synergistic. Although significantly reduced by sodium azide (P < 0-001), this synergy was not totally neutralized, and the combination of penicillin, gentamicin and sodium azide remained significantly more bactericidal than penicillin alone (P < 0-001). However, when EDTA was added to this triple combination, bactericidal activity was reduced to that of penicillin alone. The combination of EDTA and gentamicin was markedly synergistic but this synergy was almost completely abolished by the addition of sodium azide (P < 0-001). Post-antibiotic effects The strain was exposed transiently (1 h) to penicillin, gentamicin, EDTA and sodium azide, alone and in various combinations. Exposure to gentamicin, EDTA or sodium azide induced a PAE of less than 1 h; exposure to gentamicin or EDTA in the presence of sodium azide did not alter the duration of this PAE (Table I; Figure 2). The PAE induced by penicillin was longer than 2 h and was largely unaffected by addition of sodium azide (Table I, Figure 2). The PAE induced by exposure to penicillin plus gentamicin was more than double that induced by exposure to penicillin alone (Figure 3(a)). The addition of sodium azide to penicillin plus gentamicin reduced the duration of the PAE, but this still remained longer than the PAE induced by penicillin alone (Table I; Figure 3(a)). The results were similar when PAEs were determined by viable counting methods (results not shown). Monitoring of ATP leakage showed a similar pattern; leakage of ATP was enhanced
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Statistical analyses
T. G. Winstaoky and J. G. M Hastings
554
t.
13
-2
-3 10 12 -4
-5
-6 Figure 1. Effect of various compounds on enterococci. Change in viable count (cfu/ml ± S.D.) after exposure for 4 h to: (1) ISB (control), (2) sodium azide, (3) penicillin, (4) penicillin plus sodium azide, (5) gentamicin, (6) gentamicin plus sodium azide, (7) EDTA, (8) EDTA plus sodium azide, (9) penicillin plus gentamkan, (10) penicillin plus gentamicin phis sodium azide, (11) penicillin plus gentamicin plus EDTA plus sodium azide, (12) gentamicin plus EDTA, (13) gentamicin plus EDTA plus sodium azide. The concentrations used were as in Table I.
and prolonged following exposure to penicillin plus gentamicin compared with exposure to penicillin alone. This phenomenon was partially blocked by sodium azide (Figure 3(b)). PAE experiments were repeated with sequential rather than simultaneous exposure to antibiotics, sodium azide and EDTA (see Methods). Exposure of strain 281 to penicillin for 1 h followed by a further exposure for 1 h to sodium azide, EDTA or to a combination of the two resulted in a PAE of around 2-0 h (Figure 4), a similar duration to that obtained after simultaneous exposure to these agents. Exposure of cells to penicillin, then gentamicin, extended the PAE to 4-7 h. This synergy was reduced to 3-7 h when sodium azide was added with the gentamicin. However, addition of both EDTA and sodium azide to the gentamicin reduced the PAE to the same duration as that induced by penicillin alone. The combination of EDTA plus gentamicin induced a marked PAE (2-4 h) which
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
-I
555
synergy
Table L PAEs following exposure to antibiotics and EDTA, singly and in combination, in the presence and absence of sodium azide (Strain 281)
Sodium azide (025%) Absent Present
Penicillin (10 mg/1)
06 07
0-5 09
5-9 3-3
6-3 2-2
EDTA (1500 mg/1) and Gentamicin (10 mg/1) 2-4 01
was reversed by exposure to the two agents in the presence of sodium azide (Figure 5(a)). Parallel measurements of extracellular ATP showed leakage of the nucleotide during the PAE induced by EDTA plus gentamicin. Once again, sodium azide reversed this phenomenon (Figure 5(b)). Discussion It is generally accepted that, before major aminoglycoside uptake can occur, a few molecules must first enter the cell through imperfections in the growing membrane. One hypothesis is that mis-translated proteins then accumulate in the cytoplasmic membrane (Davis, Chen & Tai, 1986) allowing positively-charged molecules to diffuse inwards through transmembrane aqueous channels in response to the protonmotive force (PMF). An alternative explanation is that mistranslation gives rise to 'rogue' transport systems comprising a cycle involving a carrier-based mechanism, possibly
i
o
Time offer removal (h)
Figure 2. Post-treatment recovery of enterococci following exposure for 1 h to: (1) ISB (control), (2) sodium azide, (3) gentamicin, (4) gentamicin plus sodium azide, (5) EDTA, (6) EDTA plus sodium azide, (7) penicillin plus sodium azide, (8) penicillin. The concentrations used were as in Table I.
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
2-4 2-2
PAE (h) following exposure to: Penicillin (10 mg/1) Penicillin Gentamicin (10 mg/1) (10 mg/1) and Gentamicin Gentamicin EDTA EDTA (10 mg/1) (1500 mg/1) (10mg/I) (1500 mg/1)
T. G. Wfastanky and J. G. M. Hastings
556
(b)
3 Time after removal
4 (h)
Flgnre 3. Post-treatment recovery of enterococci following exposure for 1 h to: (1) ISB (control), (2) gentamicin plus sodium azidc, (3) penicillin plus sodium azkle, (4) penicillin plus gentamkin phis sodium azkle, (5) penicillin plus gentamicin. The concentrations used were as in Table I; (a) total ATP and (b) extracellular ATP.
coupled to an oxidation-reduction reaction (Nichols, 1987). Entcrococci arc relatively resistant to aminoglycosides (Mates et al., 1982). Nonetheless, as expected, we found that the combination of penicillin and gentamicin showed bactericidal synergy against enterococci. This may be due, in part, to the disruptive effect of penicillin on the cell wall. However, an additional effect at the bacterial membrane may facilitate the initial uptake of aminoglycoside molecules necessary for the misreading of membrane proteins, and this may be a precursor to gross aminoglycoside uptake. There is good evidence for the action of penicillin on bacterial membranes (Gale & Taylor, 1947; Prestidge & Pardee, 1957; Home, Hakenbeck & Tomasz, 1977; Brissette, Schockman & Pieringer, 1982). Mates et al. (1982) demonstrated a linear dependence of gentamicin uptake, and bacterial killing, on the magnitude of the cytoplasmic membrane potential and showed that nigericin, an ionophore that raises membrane potential, promoted uptake of gentamicin in a manner similar to that of penicillin. We have shown previously that penicillin increases the permeability of enterococci to aminoglycosides and that the susceptibility of penicillintreated cells to gentamicin correlates with the leakage of adenine nucleotides (Winstanley & Hastings, 1989). Penicillin-induced aminoglycoside uptake may be autocatalytic in the sense that aminoglycosides potentiate the membrane damage (Anand & Davis, 1961; Dubin, Hancock & Davis, 1963). Gentamicin also prolongs ATP leakage
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
•
Paridffin-gentamidii synergy
1
2
3 4 Time after removal (h)
Fignc 4. Post-treatment recovery of enterococci after exposure for 1 h to penicillin and a further 1 h exposure to: (2) sodium aide, (3) gentamicin plus EDTA plus sodium azide, (4) gentamicin plus sodium aide, (5) gentamicin. The concentrations used were as in Table I. Untreated cells (1) were used as a control.
from penicillin-treated cells (Winstanley & Hastings, 1989). In the present study, as in others (Yee et al., 1986; Fuursted, 1989), sodium azide, was used, and assumed specifically to inhibit energy-dependent steps in aminoglycoside uptake. Yee et al. (1986), stated that sodium azide prevented an increase in radioactive streptomyin uptake by viridans streptococci that had been pretreated with penicillin but did not present evidence for this contention. They suggested that penicillin has a direct effect on the membrane potential, leading to the stimulation of active transport of aminoglycosides. Miller et al. (1986), by a similar method, found that penicillin increased streptomycin uptake by enterococci but not by viridans streptococci. Stimulation of streptomycin uptake in streptococci by penicillin may be strain or species specific. Fuursted (1989) used viable counts to show bactericidal and PAE synergy between ampicillin and streptomycin against enterococci, and claimed that aminoglycoside uptake by ampicillin-damaged enterococci was a passive process because the synergy was not reversed by sodium azide or anaerobic conditions. However, the data presented by this author clearly show partial inhibition of ampicillin-streptomycin synergy by sodium azide, results similar to our own findings with penicillin and gentamicin. We also found that the PAEs induced by penicillin and gentamicin were longer than those induced by penicillin alone and that these were again partially reduced by sodium azide. Experiments to monitor bacterial leakage of ATP during PAE periods further supported these observations. The present findings are consistent with the principles of membrane energisation. According to Mitchell's chemiosmotic hypothesis a PMF may be generated by two routes in facultative anaerobic bacteria; either from aerobic respiration via electron transport or by hydrolysis of ATP generated from glycolysis (Harold, 1977; Mitchell, 1979). A possible explanation of our observations is as follows. Enterococci have a low
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
0
557
T. G. Wtostanky and J. G. M. Hastings
558
/5
f
10
3
/
(a)
/
/
/
•» "o
|
o
u
If 1/ m i
1
i
2
i
3 4 Time after removal (h)
Figure 5. Post-treatment recovery of enterococci following exposure for 1 h to: (1) ISB (control), (2) EDTA plus sodium azide, (3) gentamidn plus sodium azide, (4) gentamicin plus EDTA plus sodium azide, (5) gentamicin plus EDTA. The concentrations used were as in Table I; (a) total ATP and (b) extracellular ATP.
electrical potential across their membranes and are, therefore, relatively resistant to aminoglycoside entry. Penicillin raises the membrane potential (Mates et al., 1982), facilitating aminoglycoside entry which, in turn, leads to mistranslation events and further aminoglycoside uptake in response to the PMF. Sodium azide is a respiratory poison which blocks electron transport. The partial inhibition of penicillin-aminoglycoside synergy by sodium azide is presumably a reflection of the continued, but reduced, PMF generation via ATP hydrolysis by membrane-bound ATPase. As might be predicted, the additional presence of EDTA, which chelates divalent cations and blocks the action of ATPase (Davis & Iannetta, 1972; Bryan & Van den Elzen, 1977), leads to complete inhibition of peniciUin-aminoglycoside synergy. A possible alternative explanation might be that the sodium azide slowed cell division and thus interfered with penicillin-induced aminoglycoside uptake. However, this would seem unlikely, as sequential addition PAE experiments gave essentially similar results. The marked bactericidal and PAE synergy between EDTA and gentamicin may be a result of chelation of divalent cations at the membrane (Davis & Iannetta, 1972). Synergy was completely abolished by sodium azide. This may be a reflection of blockage of both mechanisms for PMF generation—electron transport by sodium azide and membranebound ATPase by EDTA.
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
1 ATP
o.
Penkfflin-gentamidn synergy
559
We consider that these data provide further evidence that aminoglycoside uptake is energy-dependent even when enterococci have been pre-treated with penicillin. Acknowledgements
We thank Dr D. M. Harris, Dr R. C. Spencer and Prof. B. I. Duerden for critical reading of this manuscript. We also thank K. M. Oxley for preparation of the figures. References
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Anand, N. & Davis, B. D. (1961). Damage by streptomycin to the cell membrane of Escherichia coli. Nature 185, 22-3. Anand, N., Davis, B. D. & Armitage, A. K. (1961). Uptake of streptomycin by Escherichia coli. Nature 185, 23-4. Brissette, J. L., Shockman, G. D. & Pieringer, R. A. (1982). Effects of penicillin on synthesis and excretion of lipid and lipoteichoic acid from Streptococcus mutatis BHT. Journal of Bacteriology 151, 838-44. Bryan, L. E. & Van den Elzen, H. M. (1977). Effects of membrane-energy mutations and cations on streptomycin and gentamicin accumulation by bacteria: a model for entry of streptomycin and gentamicin in susceptible and resistant bacteria. Antimicrobial Agents and Chemotherapy 12, 163-77. Davis, B. D., Chen, L. & Tai, P. C. (1986). Misread protein creates membrane channels: an essential step in the bactericidal action of aminoglycosides. Proceedings of the National Academy of Sciences of the United States of America 83, 6164-8. Davis, S. D. & Iannetta, A. (1972). Influence of serum and calcium on the bactericidal activity of gentamicin and carbcnicillin on Pseudomonas aeruginosa. Applied Microbiology 23, 775-9. Dubin, D. T., Hancock, R. & Davis, B. D. (1963). The sequence of some effects of streptomycin in Escherichia coli. Biochimica et Biophysica Ada 74, 476-89. Fuursted, K. (1989). Synergism and mechanism of subinhibitory concentration of streptomycin on Streptococcus faecalis. APMIS 97, 27-32. Gale, E. F. & Taylor, E. S. (1947). The assimilation of amino-acids by bacteria. 5. The action of penicillin in preventing the assimilation of glutamic acid by Staphylococcus aureus. Journal of General Microbiology 1, 314-26. Hancock, R. (1962). Uptake of l4C-streptomycin by some micro-organisms and its relation to their streptomycin sensitivity. Journal of General Microbiology 28, 493-501. Harold, F. M. (1977). Membranes and energy transduction in bacteria. Current Topics in Bioenergetics 6, 83-149. Home, D., Hakenbeck, R. & Tomasz, A. (1977). Secretion of lipids induced by inhibition of peptidoglycan synthesis in streptococci. Journal of Bacteriology 132, 704-17. Hunter, T. H. (1947). Use of streptomycin in treatment of bacterial endocarditis. American Journal of Medicine 2, 436-42. Mates, S. M., Eisenberg, E. S., Mandel, L. J., Patel, L., Kaback, H. R. & Miller, M. H. (1982). Membrane potential and gentamicin uptake in Staphylococcus aureus. Proceedings of the National Academy of Sciences of the United States of America 79, 6693-7. Miles, A. A., Misra, S. S. & Irwin, J. O. (1938). The estimation of the bactericidal power of the blood. Journal of Hygiene 38, 732-49. Miller, M. H., El-Sokkary, M. A., Feinstein, S. A. & Lowy, F. D. (1986). Penicillin-induced effects on streptomycin uptake and early bactericidal activity differ in viridans group and enterococcal streptococci. Antimicrobial Agents and Chemotherapy 30, 763-8. Mitchell, P. (1979). Keilin's respiratory chain concept and its chemiosmotic consequences. Science 206, 1148-59. Moellering, R. C. & Weinberg, A. N. (1971). Studies on antibiotic synergism against enterococci. II. Effect of various antibiotics on the uptake of '*C-labeled streptomycin by enterococci. Journal of Clinical Investigation 50, 2580-4. Nichols, W. W. (1987). On the mechanism of translocation of dihydrostreptomycin across the bacterial cytoplasmic membrane. Biochimica et Biophysica Acta 895, 11-23.
560
T. G. Wfastanky and J. G. M. Hastings
Plotz, P. H. & Davis, B. D. (1962). Syncrgism between streptomycin and penicillin: a proposed mechanism. Science 135, 1067-8. Prestidge, L. S. & Pardec, A. B. (1957). Induction of bacterial lysis by penicillin. Journal of Bacteriology 74, 48-59. Winstanley, T. G. & Hastings, J. G. M. (1989). Penicillin-aminoglycoside synergy and postantibiotic effect for enterococci. Journal of Antimicrobial Chemotherapy 23, 189-99. Yee, Y., Farber, B. & Mates, S. (1986). Mechanism of penicillin-streptomycin synergy for clinical isolates of viridans streptococci. Journal of Infectious Diseases 154, 531-4. {Received 10 November 1989; accepted 4 December 1989)
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Synergy between penicillin and gentamidn against enterococci T. G. Wlnstanky* and J. G. M. Hastings6*
The role of active uptake in aminoglycoside activity against penicillin-treated enterococci was studied by viable counts and ATP determinations. Penicillin and gentamicin gave syncrgistic bactericidal and post-antibiotic effects (PAEs) which were partially reduced by sodium azide, an electron transport inhibitor, and totally blocked in the presence of both sodium azide and EDTA, which chelates divalent cations. EDTA and gentamicin showed marked synergy in both 'killing curve' and PAE experiments. This synergy was completely inhibited by sodium azide. The data indicate that the activity of gentamicin against enterococci that have been damaged by penicillin or EDTA is energy-dependent This is consistent with present theories of gentamicin uptake via transportation driven by a protonmotive force.
Introduction
Synergy between penicillins and aminoglycosides was first reported by Hunter (1947), and the combination of penicillin and streptomycin has been a classic example of antibiotic synergy since it was first described in relation to Escherichia coli (Anand, Davis & Armitage, 1961). Plotz & Davis (1962) demonstrated that penicillin stimulated the uptake of streptomycin by Esch. coli. Other cell wall active agents, such as vancomycin, increase the uptake of aminoglycosides by enterococci (Moellering & Weinberg, 1971). The uptake and lethal effects of streptomycin require active respiration (Hancock, 1962) and, in Esch. coli, two energy-dependent phases of streptomycin uptake, both inhibited by respiratory poisons, have been described (Bryan & Van den Elzen, 1977). The role of active transport in aminoglycoside uptake by antibiotic-damaged Gram positive organisms is less clear. Yee, Farber & Mates (1986) showed that sodium azide, an inhibitor of electron transport, could prevent the increase in streptomycin uptake seen in viridans streptococci that had been pretreated with cell wall synthesis inhibitors. However, recent work (Fuursted, 1989) has contradicted these observations, indicating that energy is not required for aminoglycoside uptake by enterococci that have been pretreated with cell wall synthesis inhibitors. The aim of the present work was to resolve these conflicting findings by studying the interaction of penicillin and gentamicin alone and in the presence of a chelating agent and a respiratory poison. •Present address: Department of Clinical Microbiology, The Queen Elizabeth Hospital, Queen Elizabeth Medical Centre, Edgbaston, Birmingham B15 2TH, UK. 551 03O5-7453/9O/WO551 + 10 S02.00/0
© 1990 The British Society for Antimicrobial Chemotherapy
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
'Department of Bacteriology, The Royal Hallamshire Hospital, Glossop Road, Sheffield, S102JF; bDepartment of Medical Microbiology, The University of Sheffield Medical School, Sheffield S102RX, UK
552
T. G. Winstanky and J. G. M. Hastings Materials and methods
Bacteria Enterococcus faecalis strain 281 (Winstanley & Hastings, 1989) was used throughout. The MICs of penicillin and gentamicin for this strain are 1 and 8 mg/1, respectively. The organism was stored in glyccrol broth at — 70°C and was subcultured on to blood agar when required. Antibacterial agents
Determination of bactericidal activity and post-antibiotic effect (PAE) Organisms were grown in Iso-Sensitest Broth supplemented with glucose (0-1% w/v) (ISB, Oxoid) to a concentration of 10*cfu/ml. This density was attained in midlogarithmic phase. Antibacterial agents were then added to the cultures to achieve the following concentrations: benzylpenicillin (10 mg/1), gentamicin (10 mg/1), EDTA (1500 mg/1) and sodium azide (BDH; 0-25% w/v). Untreated cultures were used as controls. Bactericidal activity was assessed by surface viable counts on blood agar (Miles, Misra & Irwin, 1938) immediately after addition of the various agents to organisms and after incubation for 4h at 37°C. Post-antibiotic effects (PAEs) were determined by removal of the antibacterial agents from the bacterial cells after incubation for 1 h. Cultures were centrifuged (3000 g for lOmin), washed three times, and resuspended in fresh pre-warmed broth. Controls not exposed to antimicrobial agents were treated in the same manner. Bacterial regrowth was quantitated by measurement of total bacterial ATP, extracellular ATP or from viable counts. In further experiments, PAEs were determined after exposure of organisms in a sequential manner. The cells were exposed to penicillin (10 mg/1) for 1 h, followed by centrifugation and washing, and then exposed to gentamicin (10 mg/1), sodium azide (0-25%) or EDTA (1500 mg/1) or to combinations of the three for a further 1 h before a final washing procedure and measurement of bacterial regrowth. Assay of bacterial ATP Total bacterial ATP was determined by a bioluminescence method and taken as a measure of bacterial biomass. A 200 /d sample of culture was mixed with an equal volume of extraction reagent, trichloroacetic acid (2-5%) containing 4DIM EDTA. After 2 min, 20 /xl of this extract was added to 200 /J of buffer (Tris acetate, EDTA) and 20 /J of ATP-monitoring reagent (luciferin-luciferase; Pharmacia) in the well of a white plastic microtitration tray (Amersham International, Little Chalfont, UK). Light emission was measured in a himinomcter (Amerlite Analyser). A 10-fA volume of ATP standard (0-01 mM, Pharmacia) was then added to each well and the light output was again measured. The ATP concentration in the samples was then calculated using the formula: P = IOO(R-B)/(S-R) where P was the ATP concentration in picomoles, R the test reading, S the reading after the addition of the ATP standard and B the
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Benzylpenicillin (Glaxo) and gentamicin (Roussel Laboratories) were obtained as standard pharmaceutical preparations. Fresh solutions were prepared in distilled water before each experiment.
Penkfllin-gentamldii synergy
553
background reading of the luminometer. Extracellular ATP was determined by assay of 10/J of culture without extraction; other additions being as described above. Quantitation of PAE PAE was calculated from the equation PAE = T— C where T was the time required for the ATP content of the test culture to increase by 5 picomoles or for the viable count to increase ten-fold above the value pertaining immediately after removal of antibacterial agent and C was the time required for the control culture to increase by a similar amount. Preliminary data showed this value to occur at a stage when the bacteria were in the logarithmic growth phase.
Results were expressed as the mean and standard deviation of three readings. Student's /-test was used to compare differences in bactericidal activity. Results Bactericidal effects
Strain 281 was incubated for 4 h in the presence of various combinations of penicillin, gentamicin, EDTA and sodium azide (Figure 1). EDTA alone had a bacteriostatic effect, penicillin was weakly bactericidal but gentamicin had little effect on bacterial growth. There was a moderate suppression of growth in the presence of sodium azide. The addition of sodium azide had no significant effect on the bactericidal activity of penicillin. As expected, the penicillin plus gentamicin combination was synergistic. Although significantly reduced by sodium azide (P < 0-001), this synergy was not totally neutralized, and the combination of penicillin, gentamicin and sodium azide remained significantly more bactericidal than penicillin alone (P < 0-001). However, when EDTA was added to this triple combination, bactericidal activity was reduced to that of penicillin alone. The combination of EDTA and gentamicin was markedly synergistic but this synergy was almost completely abolished by the addition of sodium azide (P < 0-001). Post-antibiotic effects The strain was exposed transiently (1 h) to penicillin, gentamicin, EDTA and sodium azide, alone and in various combinations. Exposure to gentamicin, EDTA or sodium azide induced a PAE of less than 1 h; exposure to gentamicin or EDTA in the presence of sodium azide did not alter the duration of this PAE (Table I; Figure 2). The PAE induced by penicillin was longer than 2 h and was largely unaffected by addition of sodium azide (Table I, Figure 2). The PAE induced by exposure to penicillin plus gentamicin was more than double that induced by exposure to penicillin alone (Figure 3(a)). The addition of sodium azide to penicillin plus gentamicin reduced the duration of the PAE, but this still remained longer than the PAE induced by penicillin alone (Table I; Figure 3(a)). The results were similar when PAEs were determined by viable counting methods (results not shown). Monitoring of ATP leakage showed a similar pattern; leakage of ATP was enhanced
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Statistical analyses
T. G. Winstaoky and J. G. M Hastings
554
t.
13
-2
-3 10 12 -4
-5
-6 Figure 1. Effect of various compounds on enterococci. Change in viable count (cfu/ml ± S.D.) after exposure for 4 h to: (1) ISB (control), (2) sodium azide, (3) penicillin, (4) penicillin plus sodium azide, (5) gentamicin, (6) gentamicin plus sodium azide, (7) EDTA, (8) EDTA plus sodium azide, (9) penicillin plus gentamkan, (10) penicillin plus gentamicin phis sodium azide, (11) penicillin plus gentamicin plus EDTA plus sodium azide, (12) gentamicin plus EDTA, (13) gentamicin plus EDTA plus sodium azide. The concentrations used were as in Table I.
and prolonged following exposure to penicillin plus gentamicin compared with exposure to penicillin alone. This phenomenon was partially blocked by sodium azide (Figure 3(b)). PAE experiments were repeated with sequential rather than simultaneous exposure to antibiotics, sodium azide and EDTA (see Methods). Exposure of strain 281 to penicillin for 1 h followed by a further exposure for 1 h to sodium azide, EDTA or to a combination of the two resulted in a PAE of around 2-0 h (Figure 4), a similar duration to that obtained after simultaneous exposure to these agents. Exposure of cells to penicillin, then gentamicin, extended the PAE to 4-7 h. This synergy was reduced to 3-7 h when sodium azide was added with the gentamicin. However, addition of both EDTA and sodium azide to the gentamicin reduced the PAE to the same duration as that induced by penicillin alone. The combination of EDTA plus gentamicin induced a marked PAE (2-4 h) which
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
-I
555
synergy
Table L PAEs following exposure to antibiotics and EDTA, singly and in combination, in the presence and absence of sodium azide (Strain 281)
Sodium azide (025%) Absent Present
Penicillin (10 mg/1)
06 07
0-5 09
5-9 3-3
6-3 2-2
EDTA (1500 mg/1) and Gentamicin (10 mg/1) 2-4 01
was reversed by exposure to the two agents in the presence of sodium azide (Figure 5(a)). Parallel measurements of extracellular ATP showed leakage of the nucleotide during the PAE induced by EDTA plus gentamicin. Once again, sodium azide reversed this phenomenon (Figure 5(b)). Discussion It is generally accepted that, before major aminoglycoside uptake can occur, a few molecules must first enter the cell through imperfections in the growing membrane. One hypothesis is that mis-translated proteins then accumulate in the cytoplasmic membrane (Davis, Chen & Tai, 1986) allowing positively-charged molecules to diffuse inwards through transmembrane aqueous channels in response to the protonmotive force (PMF). An alternative explanation is that mistranslation gives rise to 'rogue' transport systems comprising a cycle involving a carrier-based mechanism, possibly
i
o
Time offer removal (h)
Figure 2. Post-treatment recovery of enterococci following exposure for 1 h to: (1) ISB (control), (2) sodium azide, (3) gentamicin, (4) gentamicin plus sodium azide, (5) EDTA, (6) EDTA plus sodium azide, (7) penicillin plus sodium azide, (8) penicillin. The concentrations used were as in Table I.
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
2-4 2-2
PAE (h) following exposure to: Penicillin (10 mg/1) Penicillin Gentamicin (10 mg/1) (10 mg/1) and Gentamicin Gentamicin EDTA EDTA (10 mg/1) (1500 mg/1) (10mg/I) (1500 mg/1)
T. G. Wfastanky and J. G. M. Hastings
556
(b)
3 Time after removal
4 (h)
Flgnre 3. Post-treatment recovery of enterococci following exposure for 1 h to: (1) ISB (control), (2) gentamicin plus sodium azidc, (3) penicillin plus sodium azkle, (4) penicillin plus gentamkin phis sodium azkle, (5) penicillin plus gentamicin. The concentrations used were as in Table I; (a) total ATP and (b) extracellular ATP.
coupled to an oxidation-reduction reaction (Nichols, 1987). Entcrococci arc relatively resistant to aminoglycosides (Mates et al., 1982). Nonetheless, as expected, we found that the combination of penicillin and gentamicin showed bactericidal synergy against enterococci. This may be due, in part, to the disruptive effect of penicillin on the cell wall. However, an additional effect at the bacterial membrane may facilitate the initial uptake of aminoglycoside molecules necessary for the misreading of membrane proteins, and this may be a precursor to gross aminoglycoside uptake. There is good evidence for the action of penicillin on bacterial membranes (Gale & Taylor, 1947; Prestidge & Pardee, 1957; Home, Hakenbeck & Tomasz, 1977; Brissette, Schockman & Pieringer, 1982). Mates et al. (1982) demonstrated a linear dependence of gentamicin uptake, and bacterial killing, on the magnitude of the cytoplasmic membrane potential and showed that nigericin, an ionophore that raises membrane potential, promoted uptake of gentamicin in a manner similar to that of penicillin. We have shown previously that penicillin increases the permeability of enterococci to aminoglycosides and that the susceptibility of penicillintreated cells to gentamicin correlates with the leakage of adenine nucleotides (Winstanley & Hastings, 1989). Penicillin-induced aminoglycoside uptake may be autocatalytic in the sense that aminoglycosides potentiate the membrane damage (Anand & Davis, 1961; Dubin, Hancock & Davis, 1963). Gentamicin also prolongs ATP leakage
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
•
Paridffin-gentamidii synergy
1
2
3 4 Time after removal (h)
Fignc 4. Post-treatment recovery of enterococci after exposure for 1 h to penicillin and a further 1 h exposure to: (2) sodium aide, (3) gentamicin plus EDTA plus sodium azide, (4) gentamicin plus sodium aide, (5) gentamicin. The concentrations used were as in Table I. Untreated cells (1) were used as a control.
from penicillin-treated cells (Winstanley & Hastings, 1989). In the present study, as in others (Yee et al., 1986; Fuursted, 1989), sodium azide, was used, and assumed specifically to inhibit energy-dependent steps in aminoglycoside uptake. Yee et al. (1986), stated that sodium azide prevented an increase in radioactive streptomyin uptake by viridans streptococci that had been pretreated with penicillin but did not present evidence for this contention. They suggested that penicillin has a direct effect on the membrane potential, leading to the stimulation of active transport of aminoglycosides. Miller et al. (1986), by a similar method, found that penicillin increased streptomycin uptake by enterococci but not by viridans streptococci. Stimulation of streptomycin uptake in streptococci by penicillin may be strain or species specific. Fuursted (1989) used viable counts to show bactericidal and PAE synergy between ampicillin and streptomycin against enterococci, and claimed that aminoglycoside uptake by ampicillin-damaged enterococci was a passive process because the synergy was not reversed by sodium azide or anaerobic conditions. However, the data presented by this author clearly show partial inhibition of ampicillin-streptomycin synergy by sodium azide, results similar to our own findings with penicillin and gentamicin. We also found that the PAEs induced by penicillin and gentamicin were longer than those induced by penicillin alone and that these were again partially reduced by sodium azide. Experiments to monitor bacterial leakage of ATP during PAE periods further supported these observations. The present findings are consistent with the principles of membrane energisation. According to Mitchell's chemiosmotic hypothesis a PMF may be generated by two routes in facultative anaerobic bacteria; either from aerobic respiration via electron transport or by hydrolysis of ATP generated from glycolysis (Harold, 1977; Mitchell, 1979). A possible explanation of our observations is as follows. Enterococci have a low
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
0
557
T. G. Wtostanky and J. G. M. Hastings
558
/5
f
10
3
/
(a)
/
/
/
•» "o
|
o
u
If 1/ m i
1
i
2
i
3 4 Time after removal (h)
Figure 5. Post-treatment recovery of enterococci following exposure for 1 h to: (1) ISB (control), (2) EDTA plus sodium azide, (3) gentamidn plus sodium azide, (4) gentamicin plus EDTA plus sodium azide, (5) gentamicin plus EDTA. The concentrations used were as in Table I; (a) total ATP and (b) extracellular ATP.
electrical potential across their membranes and are, therefore, relatively resistant to aminoglycoside entry. Penicillin raises the membrane potential (Mates et al., 1982), facilitating aminoglycoside entry which, in turn, leads to mistranslation events and further aminoglycoside uptake in response to the PMF. Sodium azide is a respiratory poison which blocks electron transport. The partial inhibition of penicillin-aminoglycoside synergy by sodium azide is presumably a reflection of the continued, but reduced, PMF generation via ATP hydrolysis by membrane-bound ATPase. As might be predicted, the additional presence of EDTA, which chelates divalent cations and blocks the action of ATPase (Davis & Iannetta, 1972; Bryan & Van den Elzen, 1977), leads to complete inhibition of peniciUin-aminoglycoside synergy. A possible alternative explanation might be that the sodium azide slowed cell division and thus interfered with penicillin-induced aminoglycoside uptake. However, this would seem unlikely, as sequential addition PAE experiments gave essentially similar results. The marked bactericidal and PAE synergy between EDTA and gentamicin may be a result of chelation of divalent cations at the membrane (Davis & Iannetta, 1972). Synergy was completely abolished by sodium azide. This may be a reflection of blockage of both mechanisms for PMF generation—electron transport by sodium azide and membranebound ATPase by EDTA.
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
1 ATP
o.
Penkfflin-gentamidn synergy
559
We consider that these data provide further evidence that aminoglycoside uptake is energy-dependent even when enterococci have been pre-treated with penicillin. Acknowledgements
We thank Dr D. M. Harris, Dr R. C. Spencer and Prof. B. I. Duerden for critical reading of this manuscript. We also thank K. M. Oxley for preparation of the figures. References
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
Anand, N. & Davis, B. D. (1961). Damage by streptomycin to the cell membrane of Escherichia coli. Nature 185, 22-3. Anand, N., Davis, B. D. & Armitage, A. K. (1961). Uptake of streptomycin by Escherichia coli. Nature 185, 23-4. Brissette, J. L., Shockman, G. D. & Pieringer, R. A. (1982). Effects of penicillin on synthesis and excretion of lipid and lipoteichoic acid from Streptococcus mutatis BHT. Journal of Bacteriology 151, 838-44. Bryan, L. E. & Van den Elzen, H. M. (1977). Effects of membrane-energy mutations and cations on streptomycin and gentamicin accumulation by bacteria: a model for entry of streptomycin and gentamicin in susceptible and resistant bacteria. Antimicrobial Agents and Chemotherapy 12, 163-77. Davis, B. D., Chen, L. & Tai, P. C. (1986). Misread protein creates membrane channels: an essential step in the bactericidal action of aminoglycosides. Proceedings of the National Academy of Sciences of the United States of America 83, 6164-8. Davis, S. D. & Iannetta, A. (1972). Influence of serum and calcium on the bactericidal activity of gentamicin and carbcnicillin on Pseudomonas aeruginosa. Applied Microbiology 23, 775-9. Dubin, D. T., Hancock, R. & Davis, B. D. (1963). The sequence of some effects of streptomycin in Escherichia coli. Biochimica et Biophysica Ada 74, 476-89. Fuursted, K. (1989). Synergism and mechanism of subinhibitory concentration of streptomycin on Streptococcus faecalis. APMIS 97, 27-32. Gale, E. F. & Taylor, E. S. (1947). The assimilation of amino-acids by bacteria. 5. The action of penicillin in preventing the assimilation of glutamic acid by Staphylococcus aureus. Journal of General Microbiology 1, 314-26. Hancock, R. (1962). Uptake of l4C-streptomycin by some micro-organisms and its relation to their streptomycin sensitivity. Journal of General Microbiology 28, 493-501. Harold, F. M. (1977). Membranes and energy transduction in bacteria. Current Topics in Bioenergetics 6, 83-149. Home, D., Hakenbeck, R. & Tomasz, A. (1977). Secretion of lipids induced by inhibition of peptidoglycan synthesis in streptococci. Journal of Bacteriology 132, 704-17. Hunter, T. H. (1947). Use of streptomycin in treatment of bacterial endocarditis. American Journal of Medicine 2, 436-42. Mates, S. M., Eisenberg, E. S., Mandel, L. J., Patel, L., Kaback, H. R. & Miller, M. H. (1982). Membrane potential and gentamicin uptake in Staphylococcus aureus. Proceedings of the National Academy of Sciences of the United States of America 79, 6693-7. Miles, A. A., Misra, S. S. & Irwin, J. O. (1938). The estimation of the bactericidal power of the blood. Journal of Hygiene 38, 732-49. Miller, M. H., El-Sokkary, M. A., Feinstein, S. A. & Lowy, F. D. (1986). Penicillin-induced effects on streptomycin uptake and early bactericidal activity differ in viridans group and enterococcal streptococci. Antimicrobial Agents and Chemotherapy 30, 763-8. Mitchell, P. (1979). Keilin's respiratory chain concept and its chemiosmotic consequences. Science 206, 1148-59. Moellering, R. C. & Weinberg, A. N. (1971). Studies on antibiotic synergism against enterococci. II. Effect of various antibiotics on the uptake of '*C-labeled streptomycin by enterococci. Journal of Clinical Investigation 50, 2580-4. Nichols, W. W. (1987). On the mechanism of translocation of dihydrostreptomycin across the bacterial cytoplasmic membrane. Biochimica et Biophysica Acta 895, 11-23.
560
T. G. Wfastanky and J. G. M. Hastings
Plotz, P. H. & Davis, B. D. (1962). Syncrgism between streptomycin and penicillin: a proposed mechanism. Science 135, 1067-8. Prestidge, L. S. & Pardec, A. B. (1957). Induction of bacterial lysis by penicillin. Journal of Bacteriology 74, 48-59. Winstanley, T. G. & Hastings, J. G. M. (1989). Penicillin-aminoglycoside synergy and postantibiotic effect for enterococci. Journal of Antimicrobial Chemotherapy 23, 189-99. Yee, Y., Farber, B. & Mates, S. (1986). Mechanism of penicillin-streptomycin synergy for clinical isolates of viridans streptococci. Journal of Infectious Diseases 154, 531-4. {Received 10 November 1989; accepted 4 December 1989)
Downloaded from http://jac.oxfordjournals.org/ at University of Michigan on January 5, 2015
