Prevalence and mechanism of resistance to 'third-generation' cephalosporins in clinically relevant isolates of Enterobacteriaceae from 43 hospitals in the UK, 1990-1991.

In a UK survey of the occurrence of extended spectrum beta-lactamases, 96 hospitals submitted a total of 3951 non-selected, non-duplicate isolates of Enterobacteriaceae from 100 patients in each hospital, 206 of these cultures being mixed and, therefore, discarded. These isolates were initially screened for strains likely to produce extended-spectrum beta-lactamases (ESBLs) by MIC determination of beta-lactams followed by a bioassay, then disc approximation test and isoelectric focusing (IEF). Isolates were further examined using two pairs of PCR primers for both blaTEM and blaSHV genes. The ability of isolates to transfer resistance to both cefotaxime and ceftazidime by conjugation and transformation were examined. Four hundred and nine cefotaxime/ceftazidime-resistant isolates (10.9%) were identified from the 3745 submitted isolates, of which 338 (9.0%) were Enterobacteriaceae, 29 Escherichia coli, 35 Klebsiella spp. and seven Hafnia alveii. IEF suggested that 17 isolates produced an ESBL, which was confirmed in most cases by PCR and hydrolysis, five isolates produced an SHV enzyme by IEF, but not confirmed by PCR, and 11 had isoelectric points in the range 8-9 suggesting a possible AmpC enzyme. Only two isolates transferred the determinants. In the case of the Klebsiella spp., 19 of the 24 ceftazidime-resistant/clavulanate-sensitive isolates were positive by PCR for a blaSHV gene. No isolates were identified as carrying blaTEM, although eight isolates had isoelectric points of 5-6.3, suggesting the presence of a possible TEM beta-lactamase. The results for the H. alveii isolates suggest that either an AmpC-like enzyme or a transferable beta-lactamase which is not TEM/SHV is present. This study shows that a wide range of genotypically and phenotypically different isolates of Enterobacteriaceae producing ESBL-like enzymes is present throughout the UK at a frequency of about 1% of unselected isolates. It is important that surveillance of resistance to these clinically important antibiotics is maintained as the occurrence of localized or more widespread outbreaks caused by bacteria producing ESBLs is to be expected.

Bactericidal activities of five quinolones for Escherichia coli strains with mutations in genes encoding the SOS response or cell division.

The bactericidal effects of five quinolones (at the optimum bactericidal concentration for strain AB1157) on 15 strains of Escherichia coli with mutations in genes for the SOS response or cell division was studied by a viable-count method. The kill rate data were normalized for growth rate and compared to those for the wild type, AB1157. Similar MICs of enoxacin and fleroxacin were obtained for all mutants; however, different mutants had differing susceptibilities to ciprofloxacin, norfloxacin, and nalidixic acid. Killing kinetic studies showed that mutants with constitutive RecA expression (recA730 and spr-55 mutants) survived longer than AB1157 with all quinolones. Mutants deficient in SOS induction, e.g., recA430 and lexA3 mutants, also survived longer, suggesting that induction of the SOS response by quinolones is harmful to wild-type cells. Recombination repair-deficient mutants (recB21, recC22, and recD1009 mutants) were killed more rapidly than AB1157, as were excision repair mutants, except with nalidixic acid. Mutants which were unable to filament (sfiA11 and sfiB114 mutants) survived longer than AB1157 with all agents, but a mutant defective in the Lon protease was killed more quickly. It was concluded that (i) recombination and excision repair were involved in the repair of quinolone-damaged DNA and (ii) continuous induction (in response to exposure to quinolones) of the SOS response, and hence induction of the cell division inhibitor SfiA, causes cell filamentation and thereby contributes to the bactericidal activity of quinolones.

Phenotypic characterization of quinolone-resistant mutants of Enterobacteriaceae selected from wild type, gyrA type and multiply-resistant (marA) type strains.

The NCTC type strains of Escherichia coli, Enterobacter cloacae, Serratia marcescens and Klebsiella pneumoniae were exposed to 3, 5 and 10 x MIC of nalidixic acid, enoxacin, ciprofloxacin, PD 117596 and PD 127391. From each strain a mutant with a high MIC of quinolones alone (gyrA) and a mutant with intermediate resistance to quinolones, some beta-lactams, chloramphenicol and tetracycline (multiply resistant, m-r) were selected on agar containing antibiotics. The gyrA mutants required a higher concentration of quinolone to inhibit DNA synthesis by 50% but quinolone uptake kinetics and outer membrane profile were the same as the wild type. The m-r mutants had similar DNA synthesis IC50 as the wild type, decreased quinolone uptake kinetics and had decreased expression of an OMP of approximately 40 kD. The gyrA and m-r mutants were then exposed to 3, 5 and 10 x MIC of the same quinolones and new mutants (F2) selected. The F2 mutants from the gyrA mutants displayed a further increase in quinolone MIC; the F2 mutants from the m-r mutants had several phenotypes: high quinolone MICs with cross resistance to other agents, high quinolone resistance alone, or intermediate quinolone resistance alone. Most F2 mutants had MICs above the recommended breakpoint concentrations for quinolones. The F2 mutants often had altered biochemical profiles (API 20E), however, only in the case of E. cloacae did this affect speciation with the strains being identified as Rhanella aquatalis.

Correlation of quinolone MIC and inhibition of DNA, RNA, and protein synthesis and induction of the SOS response in Escherichia coli.

The effects of nalidixic acid and four fluoroquinolones on DNA, RNA, and protein synthesis in the presence and absence of 20 mg of chloramphenicol per liter were examined by comparing the killing kinetics, MIC, morphological response, and maximum concentration to induce recA in Escherichia coli. All agents demonstrated paradoxical killing kinetics, in that above an optimum concentration the rate of bactericidal action was slower. Filamentation of E. coli AB1157 was observed with all quinolones up to the optimum bactericidal concentration. Addition of chloramphenicol reduced the bactericidal activity, inhibited filamentation, and abolished recA induction, but it had no effect on DNA synthesis inhibition by any of the agents. Excellent correlation was obtained between the concentration required to inhibit DNA synthesis by 50%, the MIC, the maximum concentration to induce recA, and the optimum bactericidal concentration. Evidence from this study and previously published data suggest that the primary mechanism of action of quinolones is independent of the SOS response and does not require active protein synthesis; however, induction of recA and SOS responses is consequential and enhances cell death.

The effect of mutations in the SOS response on the kinetics of quinolone killing.

The SOS response is induced in Escherichia coli by agents that damage DNA, such as quinolone antibiotics. It has been proposed that induction of the SOS response by these agents may have a role in the mechanism of quinolone action. SOS mutants derived from Escherichia coli AB1157 were investigated by susceptibility testing and killing kinetic studies at various quinolone concentrations to determine whether SOS response induction was protective or damaging to quinolone-treated bacteria. Susceptibility testing showed some differences between the SOS mutants, but killing kinetic studies demonstrated further differences, some of which could be explained with respect to the SOS phenotype. The effect of ciprofloxacin and nalidixic acid on the mutants cannot be explained with respect to the SOS phenotype, although the presence of a defective SOS response makes the bacteria less sensitive to the action of these agents. Evidence is provided that the induction of the SOS response may be protective to fleroxacin and enoxacin treated bacteria. These results suggest that quinolones may not have a common mechanism of action, as was first thought.