Mecillinam (FL 1060), a 6,3-Amidinopenicillanic Acid Derivative: Bactericidal Action and Synergy in Vitro L

Home , Mecillinam

ANTUCROBIAL AGENTS AND CHEzOTHrAPY, Sept. 1975, p. 271-276 Vol. 8, No. 3 Copyright 0 1975 American Society for Microbiology Printed in U.S.A.

Mecillinam (FL 1060), a 6,3-Amidinopenicillanic Acid Derivative: Bactericidal Action and Synergy In Vitro L. TYBRING* AND N. H. MELCHIOR Bacteriological Research Department, Leo Pharmaceutical Products, DK-2750 Ballerup, Denmark Received for publication 26 December 1974 A newly described 6d-amidinopenicillanic acid derivative, mecillinam (formerly called FL 1060), showed a high in vitro activity against Enterobacteriac- eae. The effect on Escherichia coli was bactericidal and was due to lysis of the cells. The longer the culture grew under the influence of mecillinam or the lower the inoculum, the greater the bactericidal effect. The morphology of the cells changed towards large spheric forms (2 to 5 ,m) under the influence of mecillinam. Consequently a great discrepancy between the optical density and the viable count was seen. The morphologically abnormal cells could be protected against lysis in vitro by addition of ionized compounds such as sodium chloride. Abnormal cells were more sensitive to ampicillin than normal cells. As expected synergy could be demonstrated between mecillinam and ampicillin. This was marked under experimental conditions where the abnormal cells were protected against lysis.

In a previous paper from this laboratory a new the liquid media was measured by the cryostatic group of penicillanic acid derivatives, 6,- method using a Knauer type M osmometer, and the amidinopenicillanic acids, with unusual in specific conductivity was measured as millisiemens at vitro antibacterial properties was described (4). 36 C using a conductivity meter, Radiometer type One member of CDM 2c. the group, mecillinam Bactericidal activity. The bactericidal effect of (proposed intemational nonproprietary name), mecillinam on a growing culture of Escherichia coli 6,B- [(hexahydro-lH-azepin-1-yl)-methylene- (Leo HA2) was determined by measuring the change amino ]-penicillanic acid (formerly called FL in viable count and optical density (OD) during the 1060), has been studied extensively, and it has first 6 h of incubation. Viable counts were performed been shown that this compound differs funda- in quadruplicate on NIH agar by plating 0.1 ml of mentally from the hitherto known penicillanic serial 10-fold dilutions in NIH liquid medium. The acid derivatives both in regard to antibacterial colonies were counted after 18 h at 36 C. OD was spectrum (4, 8) and its mode of action (3, 5-8). measured by means of a Vitatron MPS photometer with a tungsten light without a filter. In some The results presented in this paper deal with instances the total number of microbial cells was the bactericidal effect of mecillinam and its quantitatively determined by microscopy. synergistic action with ampicillin in vitro. In Fig. 1 to 4 the OD scale was adjusted to the viable count scale by measuring the OD of dilutions of MATERIALS AND METHODS an overnight culture of the test organism. Such a Antibiotics. Mecillinam was used as the hydro- culture consists of separate minor cells, and the chloride dihydrate and ampicillin as the trihydrate. &adjustment is therefore not entirely valid for a culture The concentrations of both compounds were calcu- in the log phase, where the cells are more voluminous lated on the basis of the anhydrous zwitterion. Solu- and many pairs or short chains occur, each represent- tions of the compounds were freshly prepared for each ing only one colony-forming unit. experiment. Antibacterial activity. Antibacterial activity was Culture media. NIH agar medium and the corre- determined by the serial dilution technique either in sponding liquid medium were used throughout the agar or fluid medium. Agar plates were simultane- study. The medium contains 0.5% yeast extract ously seeded with 12 to 14 different Enterobacteriac- (Difco), 1.5% casein hydrolysate (pancreatic), 0.1% eae from our culture collection, using a heavy inocu- dextrose, 0.25% sodium chloride, 0.005% L-cystine, lum of 3.5-,1l drops of undiluted overnight cultures. and 1% agar (Oxoid no. 1) in distilled water, pH 7.1, This inoculum produced 10-mm zones of growth on after autoclaving. In some experiments NIH agar or control plates. The inoculum in fluid medium was liquid media without sodium chloride was used as a about 104 or 106 cells per ml. The results were read basic medium to which membrane-filtered solutions after 18 h at 36 C as the minimum inhibitory concen- of different compounds were added. The osmolality of tration (MIC). 271 272 TYBRING AND MELCHIOR ANTIMICROB. AGENTS CHEMOTHER.

I.X

s~~~~ @'^*.,V.Q Mecillinam

FIG. 2. Growth of E. coli with and without 1 ,g of mecillinam per ml, recorded as viable count (VC) and OD with an inoculum of 4.9 x 106 cells per ml.

i1 2 3 i 6hrs. FIG. 1. Growth of E. coli with and without 1,gg of mecillinam per ml, recorded as viable count (VC) and OD with an inoculum of 1.2 x 106 cell per ml.

RESULTS AND DISCUSSION Bactericidal activity. The growth of E. coli (Leo HA2) was compared in broth with and without mecillinam. The changes in viable count and in cell mass (OD) were measured. Figure 1 shows the effect of 1 Ag of mecillinam per ml on an E. coli culture with an initial cell FIG. 3. Growth of E. coli uwith and without 1 Ag of count of 1.2 x 106 per ml (0.2% of an overnight mecillinam per ml, recorded as viable count (VC) and culture). The viable count declined in 6 h to less OD with an inoculum of 2.6 x 107 cells per ml. than 100 per ml. The OD, however, was only slightly less than that of the control culture for ,gm and thereafter by lysis (4, 6). Furthermore the first 2 h. After 2.5 h the OD decreased due to the viable count declined faster than the total lysis of the cells. cell count (determined by microscopy) indicat- This dicrepancy between viable count and ing an increasing number of nonviable cells. cell mass agrees with previous observations by The same low final viable count was observed us that growing cells of E. coli respond to the with various concentrations of mecillinam down action of mecillinam by first forming swollen to 0.06 ,g/ml. At a contration of 0.02 fg/ml the and later spherical cells with diameters up to 5 final count was 1,500 per ml. VOL. 8, 1975 BACTERICIDAL ACTION OF MECILLINAM 273 oxygen was introduced continously and the 11: culture was stirred, the final count increased F 15-fold, i.e., a culture with the identical inoculum as above could now accomplish 8 doublings ,IC instead of 4.3 doublings. Mecillinam (1 Ag/ml) had a marked bactericidal effect on such a culture (Fig. 4), as the initial count was reduced more than 10,000-fold in 6 h. The OD was still high after 6 h due to the remains of lysed cells. Mecillinam only exhibits its bactericidal effect when the cells are growing under the influence of the substance for relatively long periods of time. To ensure suitable growth o.Dn conditions for a dense population over a pro- Mecillinam 0 1 longed period a simple chemostat was used. A II steady-state culture containing 2 x 108 E. coli II I cells per ml responded to 1 qg of mecillinam per II ml with marked lysis. II II Table 1 summarizes the results of these II experiments and shows the correlation between I the number of cell doublings accomplished by II the control culture and the bactericidal effect of mecillinam, expressed as the reduction ratio, 0 i.e., the ratio of initial count to count after 6 h. The more cell doublings possible the higher the

0 reduction ratio at the same concentration. When the compound was added to an exponen- tially growing culture the same correlation between the reduction ratio and the number of cell doublings of the control was seen. The inoculum and the experimental condi- V.G Mecillinam tions determine the number of possible cell doublings. As the lytic effect of mecillinam depends on the time growing cells are exposed FIG. 4. Growth of E. coli with and without 1 ,g of to the compound, an apparent strong inoculum mecillinam per ml, recorded as viable count (VC) and effect can be seen. However, the inoculum effect OD, with an inoculum of 3.3 x 107 cells per ml, is not determined by the initial number of cells oxygenated, and stirred culture. but by the initial cell mass relative to that of the maximum stationary phase. Figure 2 shows the bactericidal effect of 1 yg Alterations in the composition of the medium of mecillinam per ml against 4.9 x 106 orga- influenced the results. When the 0.25% sodium nisms per ml. This inoculum gave a measurable chloride was omitted from the normal medium, OD from the beginning of the experiment. After a steep decline in viable counts occurred after 60 3 h the entire cell mass had reached a level close min instead of after 90 min. Furthermore, the to that of the maximum stationary phase. reduction ratio was 2.5 x 104 even when the Consequently the lysis and decline in viable initial count corresponded to 5% of that of an count over the next few hours were much less overnight culture. pronounced. Compared to ampicillin the bactericidal ef- When an inoculum of 2.6 x 107 cells was used, fect of mecillinam on E. coli sets in rather late. the cell mass reached the maximum stationary In a culture in normal medium with an initial phase so early that the lytic effect of mecillinam count of 1.9 x 107 cells/ml exposed to 20 ,g of was greatly diminished (Fig. 3). The final viable ampicillin per ml the reduction ratio was 1.5 x count was, however, still low compared to that 105 in 2 h. This ratio increased only to 1.8 x 105 of the control culture. in the next 4 h. The differences in mode of The initial number of cells and the experi- action of mecillinam compared to ampicillin is mental conditions limiting cell growth deter- thus reflected in the lethal effect. mine the possible number of cell doublings. The Mecillinam, in contrast to penicillins, does main factor limiting growth in unshaken cultures not inhibit murein transpeptidase, D-alanine is the exhaustion of available oxygen. When carboxypeptidase, or murein endopeptidase (5, 274 TYBRING AND MELCHIOR ANTIMICROB. AGENTS CHEMOTHER. TABLE 1. Bactericidal effect of 1 .tg of mecillinam per ml on E. coli (Leo HA2) Control culture Mecillinar Initial rection Expt viable Viable No. of culture reduction Fig. conditionsa count count cell ratio (initial no. per ml per ml doublings count to count after6h in6h atr6h a 1.2 x 106 6.4 x 108 9.1 27,000 1 a 4.9 x 106 5.1 x 108 6.7 640 2 a 2.6 x 107 5.0 x 108 4.3 0.9 3 b 3.3 x 107 8.4 x 109 8.0 12,000 4 c 3.6 x 106 4.5 x 108 6.9 51,000 c 2.4 x 107 4.6 x 108 4.3 25,000 a Experimental conditions: (a) NIH liquid medium; (b) NIH liquid medium, and oxygen introduced during stirring; and (c) NIH liquid medium without 0.25% sodium chloride.

7). Consequently no cuts in the bacterial enve- quency of mecillinam-insensitive mutants in an lope, as caused by the penicillins, are seen and E. coli culture (8), as well as the bactericidal no spheroplasts are formed during exposure to effect of the compound, depends on the sodium mecillinam (6). The murein synthesis of E. coli chloride content of the medium, further experi- is inhibited about 60% by mecillinam at concen- ments were performed. The antibacterial activ- trations from 1 to 1,000 gg/ml (7). As the cell ity of mecillinam against 12 strains of wall of Enterobacteriaceae grows by diffuse Enterobacteriaceae was determined by the agar intercalation (1), an evenly distributed weaken- plate dilution technique using various media. ing of the sacculus seems evident. These charac- The conductivity and/or the osmolality of the teristics of mecillinam and the bacteria could media were altered by addition of different account for the observed rather protracted lytic ionized or un-ionized compounds. phase, which is highly influenced by the sodium The results are summarized in Table 2. The chloride concentration of the medium. basic medium was NIH agar without sodium The partial inhibition of the murein synthesis chloride. The final concentrations of the addi- is probably insufficient to account for the entire tions as well as the osmolality and the conduc- mode of action of mecillinam. Microscopically tivity of the corresponding fluid media are given the earliest effect on growing E. coli cells is a in the table. It will be seen that lowering the loss of the normal distinct separation between osmolality and the conductivity by omitting the the phase of cell elongation and the phase of normal content of 2.5 g of sodium chloride per cross wall formation, causing development of liter from the medium did not alter the MIC. short and swollen cells. This unbalanced sepa- Increasing the concentration to 10 g/liter, how- ration is followed by incomplete segregation and ever, completely protected the E. coli (Leo loss of the normal parellel-orientated division HA2) and the Salmonella typhimurium (NCTC planes of the chromatine (6). The loss of orien- 5710) strain from the lytic action of mecillinam, tation could explain why filamentous forms, as in accordance with the results reported by produced by low penicillin concentrations, Greenwood and O'Grady (3). Microscopy never occur. The disturbance could further showed that the morphology of the protected account for the impeded, often asymmetrical, bacteria was altered towards spherical forms at ingrowth to cross wall and gradually extended concentrations of mecillinam above the MIC generation time (6), as well as the increasing found in NIH agar. Protection, as shown with loss of ability of the cell to reproduce, as de- sodium chloride, could also be seen with potas- terimend by the quantitative plating method. sium chloride and sodium sulfate. Other ionized In the early phase before cell lysis influences compounds, such as sodium acetate and phos- the optical density, viable counts of the dis- phates, had the same effect. Un-ionized com- turbed cells are influenced by the media and pounds of various molecular weights did not methods. The syntheses of deoxyribonucleic protect bacteria growing on agar regardless of acid, ribonucleic acid, and protein are not the high osmolality of the medium. Experi- inhibited (6), but as the cell orientation is ments in fluid media showed the same protec- disturbed the growth into big spheric cells tive effect of sodium chloride but sucrose also seems understandable. showed a certain protection when the results Influence of the environment. As the fre- were read after 48 h. VOL. 8, 1975 BACTERICIDAL ACTION OF MECILLINAM 275 TABLE 2. Protection to the lytic action of mecillinam by various additions to basic agar medium

Addition toma Specific MIC (Pg/ml) basic medium (mOsm/kg)a conductivity E. coli S. typhimurium (g/liter) (nmS) (36 C)Y eo HA2 NCTC 5710 None 130 5.1 0.03 0.1 NaCl, 2.5h 210 10.5 0.03 0.1 NaCl, 10 460 25.1 100 100 KCl, 13 450 29.5 > 100 100 Na2SO4, 10; H20, 37 410 26.2 > 100 30 Sucrose, 100 455 3.9 0.03 0.3 Glycerol, 32 440 4.6 0.03 0.1 Ethylene glycol, 20 485 4.9 0.03 0.03 Urea, 20 460 5.0 0.03 0.1 Formamide, 15 470 5.1 0.01 0.3 Refer to the corresponding fluid medium. i.e., NIH agar medium. Greenwood and O'Grady (2) reported that the siella pneumoniae ATCC 10273, Proteus vul- lytic response of E. coli and Proteus mirabilis to garis Leo HJ2) and two nonproducers (E. coli penicillins could be manipulated by altering the Leo HA2, Salmonella typhi NCTC 5760). The osmolality of the medium. The findings re- penicillinase producers were able to inactivate ported here emphasize differences in mode of ampicillin completely in 18 h in any heavily action between mecillinam and penicillins. The inoculated tube. Under the same conditions 25 protective power of additions, in regard to the to 93% of the mecillinam remained in medium lytic response to mecillinam, was not directly without sodium chloride and 6 to 30% remained correlated with the osmolality but more likely in the other media. with the specific conductivity of the medium. E. coli and S. typhi were protected against The significance of these findings has not mecillinam by increasing the sodium chloride been confirmed in vivo. concentration. The protection was greater the Synergy with ampicillin. The different higher the inoculum. The sensitivities were mode of action of mecillinam and penicillins independent of inoculum size in basic media and the cidal activity of the antibiotics (4) without sodium chloride. K. pneumoniae and P. strongly suggested that combinations would vulgaris showed the same protection depending exert a synergistic effect. Such an effect could on the sodium chloride concentration; the effect easily be demonstrated. Other observations of the inoculum size was marked, even at low showed that stabilized strains of mecillinam- conductivity. insensitive E. coli consisted of morphologically These observations concerning the influence abnormal cells 10 times more sensitive to am- of the conductivity of the medium on the effect picillin than the original strain. of mecillinam are analogous with the observa- As a result of these observations the following tions of Greenwood and O'Grady (2), who have experiments were performed. The MIC of mecil- shown that a spectrum of osmotic suspectibility linam, ampicillin, and mixtures of the two exists within each bacterial population regard- compounds were determined against 14 strains ing the lytic response to penicillins. of Enterobacteriaceae by the serial dilution Macroscopic examinations of the tubes and technique. Macroscopic and microscopic exami- the plates where protection against mecillinam nations were performed after 18 h. The media occurred showed in some cases partial inhibi- were basic agar or fluid medium without sodium tion of growth. Microscopic examinations of the chloride, with 5 and 10 g/liter, their specific bacteria showed that when protection did not conductivities being 5.1, 15.4, and 25.1 mS, occur morphological alterations could only be respectively. found at concentrations close to the MIC. The Representative results are given in Table 3. border between normal and abnormal appear- The results comprise the MIC values of the two ance of the cells occurred at the same concen- compounds alone and a mixture of equal parts trations of mecillinam regardless of the size of of the compounds against four bacterial strains, inoculum or the conductivity of the medium. two of which are penicillinase producers (Kleb- The mecillinam concentrations could, however, 276 TYBRING AND MELCHIOR ANTIMICROB. AGENTS CHEMOTHER. TABLE 3. Synergy between mecillinam and ampicillin in vitro: dependence on inoculum size and sodium chloride contents ot the medium l-l MIC (Ag/ml) (total concn) NaCl E. coli S. typhi K. pneumoniae P. vulgaris Medium (g/liter) Leo HA2 NCTC 5760 ATCC 10273 Leo HJ2 100/ 50/ 0/ 100/ 50/ 0/ 100/ 50/ 0/ 100/ 50/ 0/ 0 50 100 0 50 100 0 50 100 0 50 100 Basic fluid medium 0 0.01 0.03 3 0.03 0.03 0.3 0. 0.3 100 0.3 0.3 10 inoculum about 5 0.03 0.1 3 0.1 0.3 1 >100 3 100 0.3 I 30 104 cells/ml 10 0.03 0.1 3 0.3 0.3 1 >100 10 30 >100 3 100 Basic fluid medium 0 0.0l 0.03 3 0.03 0.1 1 30 3 >100 30 10 > 100 inoculum about 5 0.1 0.1 3 3 0.3 1 >100 l00 >100 >100 >100 >100 l06cells/ml 10 30 0.3 3 30 0.3 1 >100 00 > 100 > 100 >100 >100 Basic agar medium 0 0.01 0.03 3 0.03 0.1 I1 > l00 100 >100 >100 30 >100 inoculum about 5 1 0.1 10 0.3 0.3 1 >100 >100 >100 >100 > l00 > l00 5 x 106 cells/cm2 10 100 1 3 10 0.3 3 >100 >100 >100 >100 >100 >100 I I__ _ ._,- a % Mecillinam/% ampicillin. be lowered by penicillinase activity of the inocu- LITERATURE CITED lum. The protective power of sodium chloride 1. Cole, R. M. 1965. Symposiur 3on the fine structure and thus manifests itself mainly by protection of the replication of bacteria and their parts. III. Bacterial abnormal swollen to spherical forms which nor- cell-wall replication followed by immunofluorescence. mally lyse. Bacteriol. Rev. 29:326-344. It was possible to demonstrate a marked 2. Greenwood, D., and F. O'Grady. 1972. The effect of osmolality on the response of Escherichia coli and synergy between mecillinam and ampicillin for Proteus mirabilis to penicillins. Br. J. Exp. Pathol. 53: all the 14 strains of Enterobacteriaceae, al- 457-464. though the experimental conditions under 3. Greenwood, D., and F. O'Grady. 1973. FL 1060: a new which it appeared differed from strain to strain beta-lactam antibiotic with novel properties. J. Clin. Pathol. 26:1-6. as can be seen from Table 3. When the condi- 4. Lund, F., and L. Tybring. 1972. 6,-amidinopenicillanic tions, determined by the size of inoculum and acids-a new group of antibiotics. Nature (London) the concentration of sodium chloride, did not New Biol. 236:135-137. protect against the action of mecillinam, the 5. Matsuhashi, S., T. Kamiryo, P. M. Blumberg, P. Linnett, E. Willoughby, and J. L. Strominger. 1974. Mechanism effect of the combination was only additive. of action and development of resistance to a new amidino However, when the conditions allowed the ab- penicillin. J. Bacteriol. 117:578-587. normal cells to resist the lytic action of mecilli- 6. Melchior, N. H., J. Blom, L. Tybring, and A. Birch-Ander- nam, a marked synergy occurred. Microscopic sen. 1973. Light and electron microscopy of the early response of Escherichia coli to a 6,8-amidinopenicillanic examinations of the cells showed that the bor- acid (FL 1060). Acta Pathol. Microbiol. Scand. Sect. B. der between normal and abnormal cells oc- 81:393-407. curred at the same concentrations of mecilli- 7. Park, J. T., and L. Burman. 1973. FL 1060, a new peni- nam whether ampicillin was present or not. The cillin with a unique mode of action. Biochem. Biophys. Res. Commun. 51:863-868. synergy thus seems to appear because the 8. Tybring, L. 1975. Mecillinam (FL 1060), a 6,6-amidino- sensitivity to ampicillin of the protected abnor- penicillanic acid derivative: in vitro evaluation. Anti- mal cells is higher than that of the normal cells. microb. Agents Chemother. 8:266-276.