No abstract

The in vitro activities of antibiotic combinations containing ciprofloxacin and either gentamicin, sisomicin, netilmicin, amikacin, or tobramycin were evaluated by checkerboard assay (agar dilution method). A total of 220 strains of Enterobacteriaceae and Pseudomonas aeruginosa (11 species, 20 strains each) were tested. Synergistic or antagonistic effects were observed in less than 1% of the tests performed; they appeared to represent method-dependent fluctuations rather than true antibiotic interactions. No significant differences among the five aminoglycosides tested were seen. Time-kill experiments performed with three representative strains of Escherichia coli and Serratia marcescens showed additive combination effects with respect to the kill rates and inhibition of bacterial regrowth. Exposure of Serratia strains to either ciprofloxacin or gentamicin before the addition of the second drug had little influence on the combination effects observed. No antagonistic drug interactions were seen in vivo when combination therapy with ciprofloxacin and gentamicin was evaluated in a model of E. coli thigh muscle infection in neutropenic mice. Comparable therapeutic effects were obtained, regardless of whether the two compounds were administered simultaneously or sequentially at 1-or 2-h intervals.

No abstract

Metronidazole and its two principal oxidative metabolites were tested in vitro against 20 clinical isolates of the Bacteroides fragilis group. Both metabolites were bactericidal, and they exhibited 65 and 5%, respectively, of the inhibitory effect of metronidazole. Additive or weak synergistic effects resulted in combination with the parent compound.Metronidazole is metabolized in humans by oxidative mechanisms or by conjugation of the parent compound (6). The major metabolic product, 1-(2-hydroxyethyl)-2-hydroxymethyl-5-nitroimidazole (hydroxy metabolite), has been shown to occur in considerable amounts in the plasma after oral or intravenous administration of metronidazole (2,3,7). A second oxidative metabolite, 1-acetic acid-2-methyl-5-nitroimidazole (acid metabolite), was found predominantly in the urine, but was also identified in the plasma (3, 7). Little is known about the activity of these two metabolites against anaerobic bacteria. In bioassays performed with two strains of Clostridium perfringens and C. sporogenes, they exhibited 30% (hydroxy metabolite) and 5% (acid metabolite) of the activity of the parent compound (4).In this study the minimal inhibitory concentrations (MICs) and minimal bactericidal concentrations (MBCs) of metronidazole and its two principal oxidative metabolites were compared, using 20 clinical isolates of Bacteroides fragilis (10 strains), B. fragilis subsp. thetaiotaomicron (5 strains), and B. fragilis subsp. vulgatus (5 strains) as test organisms. The MICs were measured on Wilkins-Chalgren agar (8) by the reference dilution method proposed by Sutter et al. (5); results are summarized in Table 1. As indicated by the geometric means of the MICs, the hydroxy metabolite and the acid metabolite exhibited about 65 and 5% of the activity of the parent compound, respectively. The concentration required to inhibit 90% of the isolates was 2 ,g/ml for metronidazole and for the hydroxy metabolite and was 32 utg/ml for the acid metabolite. All of the test strains except one isolate of B. fragilis subsp. thetaiotaomicron showed a similar pattern of susceptibility.Since agar dilution does not permit MBC determination, the bactericidal action of the two metabolites was measured by microdilution technique in Schaedler broth, using an inoculum of about 106 bacteria per ml. Viable cell counts were performed at time zero and after 48 h of incubation, and the MBC was read as the lowest concentration to yield a 99.9% kill rate. Inhibition of growth was read from the same experiments; MICs differed by no more than two dilutions from the reference agar dilution meth-

E. coli cells treated with the bifunctional crosslinking reagents dimethyl malonimidate, succinimidate, adipimidate, suberimidate, and sebacinimidate served for the isolation of rod-shaped "ghosts." These ghosts proved to be crosslinked over their entire surface; i.e., a macromolecule (resistant to boiling 1% Na dodecyl sulfate) the size of the cell had been created. Also, ghosts could similarly be crosslinked. In both cases, the final "sacs" contained about 60-70% protein, and very little or no lipopolysaccharide. When ghosts from which phospholipid had been removed were crosslinked, the covalently closed ghosts were almost pure protein; 80-90% of their dry mass was accounted for by protein. Ammonolysis of the crosslinked material (whether stemming from crosslinked cells or ghosts) showed that the same four proteins (Na dodecyl sulfate gel bands) had been crosslinked that are found in normally prepared ghosts. These observations practically exclude the hypothesis that a fluid mosaic model of membrane structure can be applied to the outer membrane of the E. coli cell envelope; rather, extensive protein-protein interactions must exist over the whole surface of this membrane. These findings are consistent with the possibility that the ghost polypeptide chains are involved in the determination of cellular shape.We have shown that rod-shaped "ghosts," which are surrounded by the outer membrane of the cell envelope, devoid of murein, and free from all cytoplasmic material except for remaining fragments of the cytoplasmic membrane, can be isolated from Escherichia coli cells (1, 2). These ghosts consist of about 25% phospholipid, 25-30% lipopolysaccharide, and 45-50% protein. We have shown that the protein of ghosts is separable into four main bands (I, II, III, and IV) in Na dodecyl sulfate-polyacrylamide gel electrophoresis (see Fig. 2). We have speculated that one or more of these polypeptide chains, i.e., presumably by their self assembly, could be the final products of the genetic information specifying cellular shape. One prediction following from this hypothesis is that protein-protein interactions should be existent over the whole cell envelope between one or more of the proteins mentioned. We show in this communication that such appears, indeed, to be the case. MATERIALS AND METHODSCells, Media, Growth Conditions, and Preparation of Ghosts. The E. coli K12 strain W945-T3282 [a diaminopimelate plus lysine auxotroph (3)] was used in the same way as described (1, 2). Ghosts were isolated following the recently described (1) procedure II. In brief, it involves treatment of cells with Triton X-100 in 40% sucrose, urea, trypsin, and finally lysozyme.Crosslinking. All diimidoesters were prepared essentially according to Davies' and Stark's (4) version of the method of McElvain and Schroeder (5), and all dinitriles were purchased from Schuchardt (Mfinchen, Germany). Whole cells for crosslinking were, after harvesting, washed once with 150 mM NaCl and once with 1 M triethanolamine, pH 8.5. They were suspended ...