386

Article

Vol. 6, No. 4

Eur. J. Clin. Microbiol., August 1987, p..386-391 0722-2211/87/04 0386-06 S 3.00/0

Comparison of the Response of Escherichia coli to Fosfomyein and Fosmidomycin Y. K a n i m o t o

] , D. Greenwood

2*

The responses of Escherichia coli to fosfomycin and fosmidomycin were investigated by continuous turbidimetric monitoring of cultures exposed to the drugs and by microscopy. The activity of both agents was potentiated by glucose-6-phosphate, suggesting that they share the inducible hexose phosphate transport system in Escherichia colt, but several differences of response were also detected: the inoculum effect was much smaller with fosfomycin than with fosmidomyein; inhibition of bacterial growth occurred much more rapidly with fosfomycin than with fosmidomycin; and fosfomycin was able to induce the formation of spheroplasts much more rapidly than fosmidomyein. Stable resistance to fosfomycin and fosmidomycin was readily induced in cultures of Escherichia coli, and some resistant variants retained susceptibility (or partial suceptibility) to the other compound. These observations suggest that although fosfomycin and fosmidomycin may be transported into Escherichia coil by a similar mechanism, the intracellular target site may be different.

Fosfomycin and fosmidomycin are both simple phosphonic acid derivatives, but are otherwise structurally different (Figure 1). Fosfomycin is a bactericidal antibiotic which has been in use for many years in certain countries, and a new salt, trometamol fosfomycin, which is highly soluble and well absorbed after oral administration has recently been developed (1, 2). Fosmidomycin, a new experimental antibiotic, is more active than fosfomycin in vitro against most gram-negative bacteria, but is not active against grampositive bacteria (3). Kahan and coworkers (4) demonstrated that fosfomycin is incorporated into bacterial cells by two alternative routes: the L-~glycerophosphate transport pathway and the inducible hexose phosphate transport system; glucose-6-phosphate potentiates the activity of fosfomycin by inducing uptake via the hexose phosphate pathway. Fosfomycin achieves a bactericidal effect by inhibition of the early stages of bacterial cell wall synthesis (4). The primary intracellular target site of fosmidomycin has not been elucidated (5). Cross-resistance between fosfomycin and fosmidomycin has been reported in laboratory mutants and in clinical isolates, although some exceptions to cross-resistance have been observed (6, 7).

1Department of Urology, Fukui Medical School, 23 Shimoaizuki, Matuoka-cho, Yoshida-gun, 910-11 Fukui, Japan. 2Department of Microbiology and PHLS Laboratory, University Hospital, Queen's Medical Centre, Nottingham NG7 2UH, UK.

In the present study we compared the responses of Escherichia coli to fosfomycin and fosmidomycin in the presence and absence of glucose-6-phosphate and investigated the properties of resistant variants by continuous turbidimetric monitoring of cultures exposed to the drugs.

Materials and Methods Drugs. Trometamol fosfomycin was provided by Zambon Farmaceutici Italy, and fosmidomycin by Fujisawa Pharmaceutical, Japan. Glucose-6-phosphate was purchased from Sigma Chemical, UK. Suitable concentrations of antibiotics were freshly prepared in sterile distilled water as required. Glucose-6-phosphate, sterilized by filtration, was added as appropriate to achieve a concentration of 25 rag/1. Organism. The strain of Escherichia coil used (laboratory code ECSA 1), was originally isolated from infected urine. Antibiotic Titrations. MICs of antibiotics were estimated by a conventional agar dilution method with a bacterialinoculum of ca. 105 CFU per spot delivered by an automatic multipoint inoculator. The medium used for antibiotic titration was Eugonbroth (BBL, USA) solidified by addition of 1% agar as previously described (8). Plates were read after overnight incubation at 37 ~ OH I

CH3-- C~H~')CH--PO3H-

OHC--N--CH2--CHz--CH2~PO3H-

F'osfomycin

Fosmidomyein

Figure 1: Structure of fosfomycin and fosmidomyein.

Vol. 6, 1987

Fosfomycin and fosmidomycin

387

Turbidimetric Studies. A twelve-channel bacterial growth monitoring device (9) was used to monitor continuously the responses of bacteria to antibiotics. Twelve tubes, each containing 20 ml Eugonbroth with or without 25 mg/1 of glucose-6-phosphate were each inoculated with one drop of an overnight broth culture of Escherichia coll. These cultures were grown at 37 ~ in the turbidimetric device; antibiotic was added at the start of the .experiment (viable count ca. 2 • 106 bacteria/ml), or when bacterial growth had raised the opacity to 10% (ca. 2 x 107 bacteria/ml), 20 % (ca. 5 • 107 bacteria/ml) and 30,% (ca. l0 B bacteria/ml) of that of a fully-grown broth culture.

ity. When fosmidomycin was added at a point o f 30 % m a x i m u m opacity, bacterial growth continued normally until the culture entered the stationary phase. Even when exposed to a concentration o f 512 mg fosmidomycin/1 ( 1 6 • MIC), no deviation from the normal growth curve was observed when the antibiotic was added at 30 % opacity (data not shown). In subsequent experiments, fosfomycin and fosmidomycin were added at an opacity level o f 20 %.

Emergence of Resistance. Cultures initially inhibited by

Escherichia coli ECSA 1 to fosfomycin and fosmido-

fosfomycin and fosmidomycin in the absence or presence of glucose-6-phosphate, but growing later in the incubation period, were tested for increased resistance by subinoculating bacteria from these cultures into fresh broth containing graded concentrations of drug and reincubating them in the turbidimetric device. The cultures were reehallenged with antibiotic at a standard point in the logarithmic growth phase (20 % maximum opacity).

mycin in the absence or presence o f glucose-6-phosphate are shown in Figure 3. In the absence o f glucose-6-phosphate the minimum concentration o f fosfomycin that caused a deviation from normal growth (minimum antibacterial concentration; MAC) was 32 mg/1 and regrowth was not seen during the 24 h period o f observation in cultures exposed to 2 5 6 m g fosfomycin/1 (MIC for this inoculum). Inhibitory concentrations of fosfomycin caused a profound and rapid fall in opacity. With fosmidomycin growth always followed the normal curve for 1 . 5 - 2 h, after which inhibitory concentrations caused a slow, prolonged fall in the opacity. The lowest concentration o f fosmidomycin to delay growth (MAC) varied between 16 and 32mg/1 in different experiments,

Light Microscopy. Specimens for examination by light microscopy were removed from antibiotic-treated cultures of

Escherichia coli at appropriate intervals. An equal volume 0 ml) of 3 % glutaraldehyde in 0.1 M sodium cacodylate containing 10 mM MgSO4 97 H2 O) was added to each sample. The samples were left in the fixative for a minimum of 1 h, after which they were centrifuged and washed three times by distilled water and then stained with carbol fuchsin.

The general form o f the turbidimetric responses o f

Results Minimum Inhibitory Concentrations In conventional MIC titrations Escherichia coli ECSA 1 was inhibited b y 16 mg o f fosfomycin/1 and 32 mg o f fosmidomycin/1 in the absence o f glucose-6-phosphate and b y 1 mg o f fosfomycin/1 and 2 mg o f fosmidomycin/1 in the presence o f glucose-6-phosphate.

Turbidimetric System The general form o f the response o f Eseherichia coli to fosfomycin was different t o that of fosmidomycin. Figure 2 shows the response o f Escherichia coli exposed to a concentration o f 64 mg/1 of fosfomycin (4 X MIC) and fosmidomycin (2 • MIC) in different phases o f growth. In the absence o f glucose-6-phosphate, cultures exposed to fosfomycin continued to grow for approximately 1 5 - 3 0 min after which there was a rapid decline o f the opacity to a residual level o f 1 5 - 2 0 %. Several hours after antibiotic addition, the opacity sometimes increased and then declined once more. This unusual response ('secondary lysis') was observed frequently with fosfomycin, but did not occur in the presence o f glucose-6-phosphate or in experiments with fosmid omycin. Exposure o f Escherichia coli to fosmidomycin resulted in a strikingly different response: following antibiotic addition the cultures continued to grow for 1 . 5 - 2 h after which there was a slow decline in opac-

Foslomycin i00

-

80 ._>-

60 ~= 40 ~T 20

0

- Con~

Fosmidomycin tO0 8O 60 ,- 40

e~ r~

o Jt"-T 2

4

i

6

i

8 Time (hours)

T

10

i

12

Figure 2: Continuous opacity records of Escherichia coli ECSA 1 exposed to fosfomycin and fosmidomycin (64 mg/l) showing the effect of antibiotic addition (arrows) at different stages in the growth cycle.

388

Y. Kanimoto et al.

Eur. J. Clin. Microbiol.

FosmIdomycln

Fosfomycin lOO

no

o

8oI

~- 60

~

o +

I J i v',

,

, ~

,

i

I

1,2,8__

128

64

40 G6P ~ 20

I

i

IX,,,,,,,,,

,

l

o

.~ 8o ~. 60

,

16

G6P ~. 40

20

0

32 : X , ,3, 2 4 6 8 10 12 14 16 1820 22 24 Time (hours)

2 4 6 i 10iz1'4i61820~2~4 Time (hours)

Figure 3: Continuous opacity records ofEscherichia coli ECSA 1 exposed to fosfomycin and fosmidomycin

in the absence ofglucose-6-phosphate (no G6P) and presence of glucose-6-phosphate (+ G6P). Antibiotic added at arrow to achieve concentrations (rag/l) shown.

Response of Parent Strain C

/,,1 /

Fosfom 'ci n no G6P C,32

Fosm idomyci n + G6P

j-

no G6P

C,lS

Fosfomycin {no G6P}

+ G6P

c

f

J c

F

Fosfomycin (+ G6P)

Fosmidomycin (no G6P}

Fosmidomycin {+ G6P)

S S

C,32

./ 1,32

~"

_/

S

Figure 4: Continuous opacity records of Escherichia coil ECSA 1 showing cross-resistance patterns of resistant variants. The first column shows the response of the parent strain to selected concentrations (mg/l) of fosfomycin or fosmidomycin in the absence and presence of glucose-6-phosphate (G6P). The cultures marked with an asterisk were then used to investigate the response of the resistant variants in the stated conditions as shown in the adjacent panels. In each case the vertical scale represents 0 - 1 0 0 % opacity and the horizontal scale represents 0 - 2 0 h. Antibiotic was added at 20 % opacity.

while the concentration that suppressed growth throughout the incubation period was always 64 mg/1. The addition of glucose-6-phosphate to the growth medium at a concentration of 25 rag/1 markedly enhanced the activity of both agents. The MAC of fosfomycin was reduced to 1 mg/1 and that of fosmidomycin was reduced to 2-4mg/1. The overall pattern of response was reproducible, but the drug concentration at which antibacterial effects were achieved varied somewhat between experiments.

Variants emerging after exposure to fosfomycin or fosmidomycin at a concentration of 32 mg/1 in the absence of glucose-6-phosphate and 16 mg/1 in the presence of glucose-6-phosphate were selected to study cross resistance. The response of the resistant variants under various conditions is shown in Figure 4. Bacteria growing after exposure to 32 mg fosfomycin/1 in the absence of glucose-6-phosphate were able to grow uninterruptedly when reexposed to concentrations of fosfomycin or fosmidomycin that inhibited the parent

Vol. 6, 1987

Fosfomycin and fosmidomycin

strain; glucose-6-phosphate was unable to restore activity to these cultures. However, variants emerging after exposure to fosmidomycin in the absence of glucose-6-phosphate retained partial susceptibility to fosfomycin (but not to fosmidomycin) and remained responsive to the potentiating effect of glucose-6phosphate. In experiments in which variants emerging in the presence of glucose-6-phosphate were examined, cross-resistance patterns 'observed were very similar irrespective of which agent had been used to induce resistance: the variants remained partially susceptible to fosfomycin, but the potentiating effect ofglucose6-phosphate was absent or very much reduced. Results obtained in conventional MIC titrations of fosfomycin and fosmidomycin against the resistant variants revealed a similar pattern of cross resistance (Table 1).

Effects o f Fosfomycin and Fosmidomyein on Morphology The lysis seen in the turbidimetric studies was accompanied by various morphological changes in the bacterial cells. Normal Escherichia coli in the logarithmic phase are short, plump bacilli with smooth cell surfaces (Figure 5a). After exposure to fosfomycin at a concentration of 2 or 4 • MIC for 30 rain two different morphological changes were observed. A few cells continued to grow but not to divide, producing filaments 3 to 4 times the length of untreated cells; however most cells underwent transformation to spheroplasts which subsequently lysed (Figure 5b). After 60 min most of the cells had succumbed to lysis. After exposure to a concentration of fosmidomycin equal to 2 or 4 • MIC, the only morphological change observed in Escherichia coli during the first 2 h was the formation of filaments which eventually developed to 5 to 10 times their original length (Figure 5c). However, most of the bacterial cells looked normal; spheroplasts were detected onty after exposure for more than 2 h (Figure 5d). After 4 h many intact cells were still present in cultures exposed to fosmidomycin, but only a few morphologically normal cells were seen in fosfomycin-treated cultures.

Discussion Although the conventionally determined MICs of fosfomycin forEscherichia cog ECSA 1 in the presence and absence of glucose-6-phosphate were similar to those of fosmidomycin, quite different responses to the two agents were revealed by continuous turbidimetric monitoring. Whereas fosfomycin induced rapid lysis irrespective of the bacterial inoculum, cultures exposed to fosmidomycin continued to grow for about 2 h before any fall in opacity occurred and there was a very pronounced inoculum effect. Fosfomycin and fosmidornycin were both potentiated by glucose-6-phosphate (which is known to induce the hexose phosphate transport system), but the general form of the response to the agents was similar whether or not glucose.6-phosphate was present. The reason for the secondary lysis sometimes observed in turbidimetric experiments with fosfomycin (Figures 2 and 3) is unknown. Since the phenomenon was never seen when glucose-6-phosphate was present it may be related to drug accumulation by another route. A similar effect is seen with another phosphonic acid antibiotic, alafosfalin (unpublished observations). Differences in response to fosfomycin and fosmidomycin observed in the turbidimetric system were accompanied by various morphological changes in the bacterial cells. Nearly all the cells exposed to 2-4 • MIC of fosfomycin developed into spheroplasts or underwent lysis within 1 h. Fosmidomycin has also been reported to induce spheroplast formation in susceptible gram-negative bacilli (5). However, in the present study spheroplasts were detectable only after exposure to 2-4 X MIC of fosmidomycin for several hours; initially, a small proportion of the bacterial population exhibited filamentation, but most of the bacteria remained morphologically normal. The eventual formation of spheroplasts in bacteria exposed to fosmidomycin indicates an effect on the bacterial cell wall, but the delay before spheroplast formation occurred suggests that the effect is not a direct interference with cell wall peptidoglycan synthesis, as occurs with fosfomycin (4). A similar delay in achieving

Table 1: MICs of fosfomycin and fosmidomycin for the parent and resistant strains of Escherichia coli ECSA 1. MICswere determined by an agar incorporation method using an inoculum of ca. l0 s CFU/spot. MIC (mg/l) Strain Parent strain RS-FOS G-6-P (-) RS-FOS G-6-P (+) RS-FMD G-6-P (-) RS-FMD G-6-P (+)

Fosfomycin Without G-6-P With G-6-P 16 > 512 64 64 128

1 > 512 64 1 128

389

Fosmidomycin Without G-6-P With G-6-P 32 256 64 128 128

RS: resistant strain; FOS: fosfomycin; FMD: fosmidomycin; G-6-P: glueose-6-phosphate.

2 256 64 2 128

390

Y. Kanimoto et al.

Eur. J. Clin. Microbiol.

Figure 5~ Light micrographs of k's'cherichia coli ECSA 1 exposed to a: no antibiotic; b: fosfomycin (2 • MIC) for 30 min; c: fosmidomycin (2 • MIC) for 2 h;d: fosmidomycin (2 • MIC) for 4 h.

an antibacterial effect is known to follow the inhibition of folate synthesis by sulphonamides. In this case the delay is attributable to the presence of preformed folate which must be depleted by distribution among progeny cells before an antibacterial effect becomes evident (10). It is tempting to speculate that a similar type of mechanism, involving a cell wall function, may account for the delay in response to fosmidomycin. Cross-resistance patterns between fosfomycin and fosmidomycin are complex. If glucose-6-phosphate was present in the growth medium, mutants which were no longer responsive to the potentiating effect were rapidly selected by either drug. These mutants

presumably lacked the hexose phosphate transport mechanism. However, mutants selected in the absence of glucose-6-phosphate did not exhibit complete cross-resistance: those selected by fosmidomycin were less resistant to fosfomycin than to fosmidomycin and vice versa. These results suggest that the two drugs either have a different site of action within Escherichia coti, or are transported into the cells by different routes in the absence of an active hexose phosphate transport pathway. Kojo et al. (6) have presented evidence that fosfomycin and fosmidomycin are both transported by the Lm-glycerophosphate route in gram-negative bacilli, and that mutants cross-resistant to both agents lack this transport pathway. We have evidence (to be published) that fosfomycin-resistant

Vol. 6, 1987

Escherichia coli with an intact hexose phosphate transport pathway can exhibit differential susceptibility to fosfomycin and fosmidomycin in the presence of glucose-6~phosphate. This evidence together with the turbidimetric and morphological data obtained in the present study support the view that these two agents act at different target sites within Escherichia coti.

References l. Ferrari, V., Bonanomi, L., Borgia, M., Lodola, E., Marco, G.: A new fosfemycin derivative with much imprevec~ bioavailability by oral route. Chemioterapia Antimicrobica 1981,4: 59-63. 2. Greenwood, D., Coyle, S., Andrew, J.: The trometamol salt of fosfomycin : microbiological evaluation. European Urology 1987, 13, Supplement 1 : 69-75. 3. Neu, H.C., Kamimura, T.: In vitro and in vivo antibacterial activity of F R-31564, a phosphonic acid antimicrobiat agent. Antimicrobial Agents and Chemotherapy 1981, 19: 1013-1023.

Fosfomycin and fosmidomycin

391

4. Kahan, F.M., Kahan, J.S., Cassidy, P.J., Kropp, H.: The mechanism of action of fosfomycin (phosphonomyein). Annals of the New York Academy of Sciences 1974, 235: 364-386. 5. Okuhara, M., Goto, T.: New phosphonic acid antibiotics produced by strains of Streptomyces. Drugs under Experimental and Clinical Research 1981, 7: 559-564. 6. Ko]o, H., Shigi, Y., Nishida, M.: FR-31564, a new phosphonic acid antibiotic: bacterial resistance and membrane permeability. Journal of Antibiotics 1980, 33: 44-48. 7. Mendez, F..I., Aivarez, A.A., Mendoza, M.C., Hardisson, C.: Antibacterial activity of fosmidomycin on chromosomic and plasmid-determined fosfomycin-resistant strains. Chemioterapm 1985, 4: 170-175. 8. Greenwood, D., Jones, A., Eley, A.: Factors influencing the activity of the trometamol salt of fosfomycin. European Journal of Clinical Microbiology 1986, 5: 29-34. 9. Mackintosh, I.P., O'Grady, F., Greenwood, D., Watson, B.W., Crichton, T.C., Piper, R., Ferret, A.: A twelve channel bacterial growth monitoring system. Biomedical Engineering 1973, 8:514-515 and 52610. Greenwood, D., O'Grady, F.: Activity and interaction of trimethoprim and sulphamethoxazole against Escherichia coll. Journal of Clinical Pathology t 9?6, 2 9 : 1 6 2 166,