ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, OCt. 1979, p.

439-443

0066-4804/79/10-0439/05$02.00/0

Vol. 16, No. 4

Mechanism of Synergistic Action of a Combination of Ampicillin and Dicloxacillin Against a ,B-Lactamase-Producing Strain of Citrobacter freundii
JUNZO MIZOGUCHI,t HIDEKAZU SUGINAKA,* AND SHOZO KOTANI Department of Microbiology and Oral Bacteriology, Osaka University Dental School, 3-48, Nakanoshima-4chome, Kita-ku, Osaka 530, Japan

Received for publication 18 July 1979

The mechanism of synergistic activity of a combination of ampicillin and dicloxacillin was studied on fl-lactamase-producing Citrobacter freundii GN346 and its derived f,-lactamaseless mutant GN346/16. The synergistic activity was exhibited against the parent strain but not against the mutant strain. Precultivation of the parent strain with the combination reduced the amount of the subsequent binding of [14C]penicillin G to the membrane fraction from the treated cells, but no reduction was observed in the case of cells treated with ampicillin or dicloxacillin alone. On the other hand, the amount of binding of [14C]penicillin G to the membrane fraction from. the mutant strain was reduced by ampicillin treatment alone. These results clearly indicated that dicloxacillin inhibited the B-lactamase activity produced by the parent strain, and, consequently, ampicillin can penetrate through the outer membrane and periplasmic ,B-lactamase barrier into its target sites on the cytoplasmic membrane.

Synergistic activity of the combination of penicillins or cephalosporins and isoxazolylpenicillins has been well known against f8-lactamaseproducing gram-negative organisms (1, 7,-8, 1214, 18, 25). No report is available on the mechanism of the synergistic effect of the combination in a whole-cell system, although isoxazolylpenicillin has been reported to inhibit soluble f8-lactamase activity in a cell-free system (1, 3, 4, 9, 12, 13). This report deals with the mechanism of the synergistic activity of a combination of ampicillin and dicloxacillin against a ,B-lactamase-producing strain of Citrobacter freundii and its derived /8-lactamaseless mutant strain, comparing the penetration of single or combined antibiotics into penicillin target sites on the cytoplasmic membrane.
MATERLALS AND METHODS Organisms. The organisms used were C. freundii

GN346 and its derived fB-lactamaseless mutant GN346/16 (19), kindly provided by S. Yamagishi (Faculty of Pharmaceutical Sciences, Chiba University, Chiba, Japan). Antibiotics. ['4C]penicillin G (labeled with [1_-4C]phenylacetic acid) was purchased from The Radit Present address: Research Laboratories, Toyo Jozo Co. Ltd., 632-1, Mifuku, Ohito-cho, Tagata-gun, Shizuoka 410-23, Japan.

ochemical Centre (Amersham, England). It had a specific activity of 52 mCi/mmol. Unlabeled penicillin G, ampicillin, and dicloxacillin were supplied by Toyo Jozo Co. Ltd. (Shizuoka, Japan). Cultivation and preparations ofthe membrane and soluble fractions. Each strain was routinely grown at 37C in Trypticase soy broth (BBL Microbiology Systems, Cockeysville, Md.) with shaking. The cells were harvested by centrifugation at 6,000 x g for 10 min at about half-maximal growth, which required approximately 3.5 h from a 5% inoculum of the overnight culture. They were thoroughly washed with 50 mM tris(hydroxymethyl)aminomethane-hydrochloride buffer, pH 7.5, containing 1 mM MgCl2 and resuspended in 1/100 of the original culture volume with the same buffer. A portion of these suspensions was used for the f,-lactamase assay of the intact cells. The suspended cells were disrupted with a Super Sonic vibrator (UR-150, Tominaga Works Ltd., Tokyo, Japan), and cell debris was removed by centrifugation at 8,000 x g for 10 min. The membrane fraction was then sedimented by centrifugation at 100,000 x g for 40 min, and the supernatant was used as the soluble fraction as described previously (23). Assay for 8-lactamase activities. /i-Lactamase activities were estimated iodometrically by a modification of Perret's method (15). The enzyme reaction was carried out at 30C in 0.1 M phosphate buffer, pH 6.8, containing 8 mM cephaloridine as a substrate. Dicloxacillin (final concentration, 1 nmol/ml) was added simultaneously to the reaction mixture for demonstration of the inhibitory effect. One unit was defined as the amount of activity capable of hydrolyzing 1 ,umol of the substrate per min under this condition.
439

440

MIZOGUCHI, SUGINAKA, AND KOTANI

ANTIMICROB. AGENTS CHEMOTHER.

Susceptibility test. Minimal inhibitory concentrations were estimated by the agar plate dilution method in heart infusion agar (Difco) with an inoculum of one loopful of 106 bacteria per ml. Assay for binding of [14Cjpenicillin G to the membrane fraction. The method used to assay ['IC] penicillin G binding was as described previously (24). The assay mixture, containing 25 1I of the membrane fraction (approximately 20 mg of protein per ml), 10 ,[l of 1 M tris(hydroxymethyl)aminomethane-hydrochloride buffer, pH 7.5, 2 .tL of 1 M MgCl2, 3 id of 10 of ['4C]penicillin G solution varying water, and lul in concentrations from 6.25 to 100 nmol/ml, was incubated for 30 min at 30C. ['4C]penicillin G bound to the membrane was separated by paper chromatography from unbound '4C-labeled compounds and counted in a liquid scintillation spectrometer (type LSC-653, Aloka Ltd., Tokyo, Japan). As a control, the binding of ['4C]penicillin G to the boiled membrane fraction (100C, 10 min) was also measured for exclusion of nonspecific binding. Pretreatment of culture celis by antibiotics. Ampicillin, dicloxacillin, or ampicillin plus dicloxacillin (2:1, by weight) was added to a final concentration of 200 tcg/ml each into the culture broth, which had been cultivated for 3.5 h at 370C. After a further cultivation for 20 min, the cells were harvested. The membrane fraction from each treated cell culture was incubated with 7.5 nmol of ['4C]penicillin G per ml in the same incubation mixture as above.
RESULTS

Synergism of a combination of ampicillin and dicloxacillin. Table 1 shows the antibacterial activities of ampicillin and dicloxacillin alone or in combination against a fl-lactamaseproducing strain, C. freundii GN346, and its derived 8-lactamaseless mutant GN346/16. The

former strain was resistant to ampicillin and to dicloxacillin, but the latter strain was susceptible to ampicillin alone. On the other hand, the former strain was susceptible to ampicillin and dicloxacillin in combination. In contrast, the combination did not exhibit synergistic activity against the fl-lactamaseless mutant strain. Effects of dicloxacillin on ,-lactamase activities. The results in Table 2 show f,-lactamase activities and their inhibition by dicloxacillin. f3-Lactamase activities of intact cells and the soluble fraction from the broken-cell preparation of C. freundii GN346/16 was reduced >500 times in comparison with the parent strain. The membrane fraction from C. freundii GN346 also showed 8-lactamase activity, whereas the activity could not be detected in the membrane fraction from the mutant strain. The /i-lactamase activities of the soluble fractions from both strains were inhibited by the addition of dicloxacillin, as had been previously reported (1), but did not destroy dicloxcillin to any measurable extent (data not shown). The activities of intact cells from both strains and the membrane fraction from the parent strain showed the same results. Binding of ['4Cjpenicillin G to the membrane fraction. When each suspension of the membrane fraction from C. freundii GN346 and its mutant strain GN346/16 was incubated with various concentrations of [14C]penicillin G for 30 min, the antibiotic was irreversibly bound to both membrane fractions (Fig. 1). The binding of ['4C]penicillin G to the membrane fraction from the mutant strain followed saturation-type kinetics, as had been shown earlier with membrane fractions from various organisms (6, 21,

TABLE 1. Minimal inhibitory concentrations of ampicillin and dicloxacillin against C. freundii GN346 and

GN346/16 Minimal inhibitory concn (#g/ml)

Strain

Ampicillin
800
1.6

Dicloxacillin
400 400

Ampicillin plus dicloxacillin

(2:1, by wt) 50 1.6

Cephaloridine
400 3.1

GN346 GN346/16

TABLE 2. Comparison of /8-lactamase activities and their inhibition by dicloxacillin in C. freundii GN346 and GN346/16 ,6-Lactamase activity
Intact cells Strain Addition Strm(U/mg of cells)

Soluble fraction
39.0 7.8 0.06 0.01

After cell disruption (U/mg of protein)

Membrane fraction
2.9 0.31
-a

GN346

None

Dicloxacillin (1 nmol/ml)
None

GN346/16
a

Dicloxacilhin (1 nmol/mi) Not detectable.

27.3 8.9 0.04 0.01

VOL. 16, 1979

SYNERGISTIC ACTION OF AMPICILLIN-DICLOXACILLIN

c 0 ^

441

-3
CD

( A)

,2
E 4

4,
In L-

E
0

m c C
1-

E .0
.0

0~

E. 5 '

0~
4-

5 10 Dir loxacillin (n moles/ml )

0 5 10 15 20 D moles/m4) 4cJ Penicillin G (n

FIG. 1. Binding of [14C]penicillin 0 to the membrane fractions from C. freundii GN346 and GN346/ 16. The membrane fractions from (A) C. freundii GN346 and (B) its derived mutant strain GN346/16 were incubated for 10 min at 30C with () or without (0) 5 nmol of dicloxacillin per ml and then were incubated with various concentrations of ['4C]penicillin G as described in the text. Data are expressed as nanomoles of bound radioactivity per milligramn of membrane proteins.

FIG. 2. Effects of dicloxacillin on binding of ['4C]penicillin G to the membrane fractions of C. freundii GN346 and GN346/16. The membrane fractions from (A) C. freundii GN346 and (B) its derived mutant strain GN346/16 were preincubated for 10 min at 30C with various concentrations of dicloxacillin, followed by the addition of a final concentration of 7.5 nmol of ['4CJpenicillin G per ml for a further incubation at 30C. Data are expressed as nanomoles of bound radioactivity per milligram of membrane proteins.

22, 24), and the saturation of penicillin-specific target sites was achieved at approximately 7.5 nmol/ml (Fig. 1B). On the other hand, the binding to the membrane fraction from the ,B-lactamase-producing strain followed nonsaturationtype kinetics (Fig. 1A), suggesting that ['4C]penicillin G in the reaction mixture was hydrolyzed by ,B-lactamase in the membrane fraction and could not bind to its specific target sites. As fl-lactamase activity of the membrane fraction from C. freundii GN346 was inhibited by dicloxacillin, as shown in Table 2, the effect of dicloxacillin on the binding of ['40]penicillin G was studied (Fig. 2). The amount bound to the membrane fraction from the parent strain increased with the addition of low concentrations of dicloxacillin, and the optimum concentration of dicloxacillin required for maximum binding under the conditions of the experiment was 5 nmol/ml (Fig. 2A). This result showed that the above concentration of dicloxacillin inhibited membrane-bound fi-lactamase activity and that ['4C]penicillin G could bind to its specific targets. On the other hand, the binding ability in

the 8i-lactamaseless mutant membrane was reduced by the addition of dicloxacillin (Fig. 2B). The binding of [14C]penicillin G to the membrane fraction from the parent strain examined after preincubation with 5 nmol of dicloxacillin per ml followed the same saturation curve as that from the mutant strain (Fig. 1A). Thus, the following experiment on the binding of [14C]_ penicillin G to the membrane fraction from the parent strain was done after preincubation with 5 nmol of dicloxaci-lin per ml to obtain the maximal amount of binding. Binding of [14C]penicillin G to the membrane fraction from penicillin-treated cels. Binding of ampicillin to its target sites was determined by measuring the reduction in [14C]penicillin G-binding to the membrane fraction prepared from cells precultivated with unlabeled ampicillin and dicloxacillin alone and in combination (Table 3). Neither ampicillin nor dicloxacillin treatment of the cells of the fl-lactamaseproducing strain, C. freundii GN346, reduced the amount of subsequent binding of ['4C]penicillin G to the membrane fraction from the treated cells. However, pretreatment with the combination of ampicillin and dicloxacillin re-

442

MIZOGUCHI, SUGINAKA, AND KOTANI

ANTIMICROB. AGENTS CHEMOTHER.

TABLE 3. Binding of ['4C]penicillin G to the membrane fractions from cells of C. freundii GN346 and GN346/ 16 treated with ampicillin and dicloxacillin alone or in combination.
Strain Treatment (200 Ag/rl)

G bound ['4C]penicilhin (dpm/mg of protein)

889 837 838 477

Penetration' (%)

GN346

None

Ampicillin Dicloxacillin Ampicillin plus dicloxacillin

(2:1, by wt)

0 5.8 5.7 46.3

GN346/16

None

Ampicilin Dicloxacillin Ampicillin plus dicloxacillin

868 331 664 350

0 61.9 23.5 59.7

(2:1, by wt) a Penetration of antibiotic to its target sites was calculated according to the following formula: (1 - [bound radioactivity to the membrane fraction from the treated cells/bound radioactivity to the membrane fraction from the nontreated cells]) x 100.

duced the binding ability. On the other hand, the amount of binding of [14C]penicillin G to the membrane fraction from the f?-lactamaseless mutant strain, C. freundii GN346/16, was reduced by treatment with ampicillin. Dicloxacillin treatment also resulted in a small extent of reduction in the binding.

DISCUSSION
A /8-lactamase-producing strain, C. freundii GN346, was resistant to ampicillin, whereas its derived f-lactamaseless mutant strain, GN346/ 16, was susceptible to this antibiotic. This indicated that ,8-lactamase of the parent strain can be considered to be an ampicillin resistance factor as suggested previously (16, 17, 26). A combination of ampicillin and dicloxacillin exhibited a marked synergistic effect against the parent strain. The mechanism of the synergistic activity has been attributed to the fact that dicloxacillin inhibited fi-lactamase activity of the parent strain, as with various f8-lactamase-producing

gram-negative organisms (1). In general, fl-lactam antibiotics are known to be irreversibly and specifically bound to the isolated membrane and inhibit transpeptidase and D-alanine carboxypeptidase activities (2,20). The present investigation revealed that the binding of ['4C]penicillin G to the membrane fraction from C. freundii GN346/16 followed the same saturation-type kinetics as those from various organisms (6, 21, 22, 24), and the binding to the membrane fraction from the parent strain was nonspecific. This result suggested that [I4C]-penicillin G in the reaction mixture was hydrolyzed by the fB-lactamase present in the membrane fraction and could not bind to its specific target sites, whereas it could bind to its specific sites when membrane-bound enzyme activity was inhibited by the addition of dicloxa-

cillin. In fact, the membrane fraction from the parent strain showed fl-lactamase activity, as reported previously in the case of a strain of Bacillus licheniformis (27), whereas the membrane fraction from the mutant strain did not show fi-lactamase activity. It would be of interest to know whether the membrane-bound filactamase is the same as the soluble enzyme present in the periplasmic space, and this is currently under study. To demonstrate the synergistic effects of the above combination on growing cells, we showed that precultivation of the cells of C. freundii GN346 with the combination of ampicillin and dicloxacillin reduced the amount of subsequent binding of ['4C]penicillin G to the membrane fraction from the treated cells, whereas no reduction was observed with precultivation with ampicillin or dicloxacillin separately. In contrast, the amount of binding of ['4C]penicillin G to the membrane from the mutant strain was reduced by ampicillin treatment alone. Thus, the present study clearly demonstrates the mechanism of synergistic action of the combination of ampicillin and dicloxacillin; dicloxacillin, a 8-lactamase inhibitor, penetrated through the outer membrane (5) and inhibited ,8-lactamase activity localized mainly in the periplasmic space (10, 11); consequently, ampicillin could penetrate through two barriers, outer membrane and periplasmic /3-lactamase, enabling the latter antibiotic to reach and bind to its specific target sites on the cytoplasmic membrane.
LfTERATURE CMD
1. Bach, J. A., N. Buono, D. Chisholm, K. E. Price, T. A. Pursiano, and A. Gourevitch. 1967. In vitro and in vivo synergism of mixtures of penicillins, p. 328-336. Antimicrob. Agents Chemother. 1966. 2. Blumberg, P. M., and J. L. Strominger. 1974. Interac-

VOL. 16, 1979

SYNERGISTIC ACTION OF AMPICILLIN-DICLOXACILLIN

443

tion of penicillin with the bacterial cell: penicillin-binding proteins and penicillin-sensitive enzymes. Bacteriol. Rev. 38:291-335. 3. Boman, H. G., K. Nordstrom, and S. Normark. 1974. Penicillin resistance in Escherichia coli K 12: synergism between penicillinases and a barrier in the outer part of the envelope. Ann. N.Y. Acad. Sci. 235:569-586. 4. Cole, M., S. Elson, and P. D. Fullbrook. 1972. Inhibition of ,f-lactamases of Escherichia coli and Klebsiella aerogenes by semi-synthetic penicillins. Biochem. J. 127:
295-308. 5. Costerton, J. W., J. M. Ingram, and K. J. Cheng. 1974. Structure and function of the cell envelope of gram-negative bacteria. Bacteriol. Rev. 38:87-110. 6. Edwards, J. R., and J. T. Park. 1969. Correlation between growth inhibition and the binding of various penicillins and cephalosporins to Staphylococcus aureus. J. Bacteriol. 99:459-462. 7. Greenwood, D., and F. 0. Grady. 1975. Potent combinations of beta-lactam antibiotics using the beta-lactamase inhibition principle. Chemotherapy 21:330-341. 8. Hamilton-Miller, J. M. T. 1971. The demonstration and significance of synergism between beta-lactam antibiotics. J. Med. Microbiol. 4:227-237. 9. Hamilton-Miller, J. M. T., and J. T. Smith. 1964. Inhibition of penicillinases from gram-positive and gram-negative bacteria by substrate analogues. Nature (London) 201:999-1001. 10. Neu, H. C., and J. Chou. 1967. Release of surface enzymes in Enterobacteriaceae by osmotic shock. J. Bacteriol. 94: 1934-1945. 11. Neu, H. C., and E. B. Winshell. 1970. Purification and characterization of penicillinases from Salmonella typhimurium and Escherichia coli. Arch. Biochem. Biophys. 139:278-290. 12. Nishida, M., and Y. Mine. 1969. Synergistic activity of ampicillin and cloxacillin. J. Antibiot. 22:144-150. 13. O'Callaghan, C. H., and A. Morris. 1972. Inhibition of f?-lactamases by ,B-lactam antibiotics. Antimicrob. Agents Chemother. 2:442-448. 14. Okubo, T., M. Inoue, and S. Mitauhashi. 1975. Antibacterial activity of combinations of cefazolin and semisynthetic penicillins. J. Antibiot. 28:804-808. 15. Perret, C. J. 1954. Iodometric assay of penicillinase. Nature (London) 174:1012-1013. 16. Richmond, M. H., and N. A. C. Curtis. 1974. The

17.
18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

interplay of /?-lactamases and intrinsic factors in the resistance of gram-negative bacteria to penicillins and cephalosporins. Ann. N.Y. Acad. Sci. 235:553-568. Richmond, M. H., and R. B. Sykes. 1973. The ,-lactamases of gram-negative bacteria and their possible physiological role. Adv. Microb. Physiol. 9:31-88. Sabath, L. D., and E. P. Abraham. 1964. Synergistic action of penicillins and cephalosporins against Pseudomonaspyocyanea. Nature (London) 204:1066-1069. Sawai, T., S. Nakajima, T. Morohoshi, and S. Yamagishi. 1977. Thermolabile repression of cephalosporinase synthesis in Citrobacter freundii. Microbiol. Immunol. 21:631-638. Strominger, J. L., P. M. Blumberg, H. Suginaka, J. Umbreit, and G. G. Wickus. 1972. How penicillin kills bacteria: progress and problems. Proc. R. Soc. London Ser. B 179:369-383. Suginaka, H. 1976. Comparison of the binding of penicillin G to staphylococcal L-form and its parent'rain membranes. Antimicrob. Agents Chemother. 9:544545. Suginaka, H., P. M. Blumberg, and J. L. Strominger. 1972. Multiple penicillin-binding components in Bacillus subtilis, Bacillus cereus, Staphylococcus aureus, and Escherichia coli. J. Biol. Chem. 247:5279-5288. Suginaka, H., A. Ichikawa, and S. Kotani. 1974. Penicillin-resistant mechanisms in Pseudomonas aeruginosa: effects of penicillin G and carbenicillin on transpeptidase and D-alanine carboxypeptidase activities. Antimicrob. Agents Chemother. 6:672-675. Suginaka, H., A. Ichikawa, and S. Kotani. 1975. Penicillin-resistant mechanisms in Pseudomonas aeruginosa: binding of penicillin to Pseudomonas aeruginosa KM 338. Antimicrob. Agents Chemother. 7:629-635. Sutherland, R., and F. R. Batchelor. 1964. Effect of penicillinase on the activity of penicillins: synergistic activity of penicillins against penicillinase-producing gram-negative bacilli. Nature (London) 201:868-869. Sykes, R. B., and M. Matthew. 1976. The f,-lactamases of gram-negative bacteria and their role in resistance to f-lactam antibiotics. J. Antimicrob. Chemother. 2:115157. Yamamoto, S., and J. 0. Lampen. 1976. Purification of plasma membrane penicilWinase from Bacillus licheniformis 749/C and comparison with exoenzyme. J. Biol. Chem. 251:4095-4101.