Search

Objectives: Fosfomycin is a possible oral treatment for lower urinary tract infections caused by Escherichia coli with CTX-M extended-spectrum beta-lactamases but is vulnerable to mutational resistance. Hypermutability among natural E. coli populations might facilitate the emergence of resistance to fosfomycin. We therefore examined the prevalence of mutators amongst urinary isolates of E. coli producing CTX-M beta-lactamases., Methods: Urinary E. coli isolates with CTX-M beta-lactamases (n = 220) were screened for resistance to both rifampicin and fosfomycin, as well as a mutator phenotype, by rifampicin and fosfomycin disc assays. Mutation frequencies for 10 isolates, identified as mutators by the initial disc screen, were determined in triplicate on agar with rifampicin or fosfomycin at 4x MIC and with fosfomycin or nitrofurantoin at 256 mg/L., Results: The disc screen identified 10 likely mutators and quantitative tests indicated that 9 of these had mutation frequencies of 8.0 x 10(-6)-1.5 x 10(-4) for fosfomycin and 0.1-2.3 x 10(-6) for rifampicin. These mutators were diverse in terms of PFGE type and 4 of the 10 were confirmed as strong mutators with rifampicin and fosfomycin. Only the strongest mutator isolate and hypermutable MutS(-) control strain consistently gave single-step mutants resistant to 256 mg/L fosfomycin. No nitrofurantoin-resistant mutants were selected from any isolate, although they could be selected from the hypermutable MutS(-) control strain., Conclusions: Mutator phenotypes were found among E. coli expressing CTX-M beta-lactamases and were independent of strain type. These had an increased propensity to fosfomycin resistance.

Objectives: To investigate the effect of subinhibitory concentrations of ciprofloxacin, streptomycin, trimethoprim, ampicillin and erythromycin on mutation frequency in Streptococcus pneumoniae., Methods: Frequency of mutation to rifampicin resistance was determined in three clinical isolates grown with or without antibiotic treatment. dinB was analysed using PCR and DNA sequence determination., Results: Subinhibitory levels of ciprofloxacin and streptomycin increased the frequency of mutation to rifampicin resistance between 2- and 5-fold for all three isolates, which is comparable to the increase seen in mismatch repair mutants of this species. These increases appeared not to be dependent on the function of the error-prone DNA polymerase encoded by dinB, since one of the isolates was a naturally occurring deletion mutant for this gene. Trimethoprim increased the mutation frequency for two isolates, but not the dinB mutant; ampicillin and erythromycin had no significant effect on mutation frequencies for any isolate., Conclusions: Exposure to quinolones and aminoglycosides at subinhibitory concentrations may result in increased mutability in pneumococci, as well as selecting for resistance per se.

Objectives: We investigated whether sulphonamide resistance in Escherichia coli remained prevalent in 2004, 9 years since the formal introduction of a UK prescribing restriction on co-trimoxazole. Resistance to other agents no longer in common use was also examined., Methods: Consecutive urinary E. coli isolates were obtained at the diagnostic microbiology laboratory of the Royal London Hospital from January to March 2004. The presence of the sulphonamide resistance genes, sul1, sul2 and sul3, and the class I integrase gene, int1, were determined by PCR., Results: Of the 391 E. coli isolates recovered in 2004, 45.5% were sulphonamide-resistant compared with 46.0% in 1999 and 39.7% in 1991. The sul2 gene remained the most prevalent sulphonamide resistance determinant, present in 81% of resistant isolates in 2004 compared with 79% and 67% in 1999 and 1991, respectively; 28% of resistant isolates carried both sul1 and sul2 genes; sul3 was not found. Resistance to streptomycin also remained common, whereas resistance to chloramphenicol and kanamycin had decreased since 1999., Conclusion: Sulphonamide resistance in E. coli persists undiminished despite the prolonged withdrawal of this antibiotic in the UK; resistance to streptomycin also seems stable whilst that to chloramphenicol and kanamycin is declining.

Objectives: Despite a 97% reduction in clinical sulphonamide usage, the prevalence of sulphonamide resistance among Escherichia coli has remained constant in the UK. Genetic linkage of sulphonamide resistance to other resistances is thought important for this maintenance, but the finding also implies that sulphonamide resistance exerts little fitness cost. To test this hypothesis, we examined the fitness impact of four naturally occurring sul2-coding plasmids upon their hosts., Methods: The fitness impact of the plasmids upon E. coli was determined by pairwise growth competition in a minimal medium. The DNA sequence of plasmid p9123 was obtained by primer walking and PCR., Results: Three of the four sul2-coding plasmids studied imposed fitness costs on their hosts. The fourth plasmid, a 6.2 kb resistance element carrying sul2, strA and strB designated p9123, conferred a 4% fitness advantage upon its original clinical host and also on E. coli K12 JM109. The complete sequence of p9123 revealed eight open reading frames, including five of unknown function. There was no obvious gene to which the fitness advantage might be attributed., Conclusions: The novel finding that p9123 can improve host fitness may explain why this plasmid and its close relatives are so widespread among enteric bacteria. In addition to other factors such as co-selection of sulphonamide resistance by other agents, the fitness advantage conferred by plasmids such as p9123 may have contributed to the maintenance of sulphonamide resistance in the UK in the absence of clinical selection pressure. These data indicate that once antibiotic resistance has been established on mobile genetic elements, it may be difficult to eliminate.

MICs of piperacillin/tazobactam are conventionally determined by varying the concentration of piperacillin in the presence of a fixed 4 mg/L tazobactam. When tested in this way, the MIC distribution for Klebsiella isolates with extended-spectrum beta-lactamases (ESBLs) is strongly bimodal, such that many producers are inhibited at 16 + 4 mg/L whilst others require MICs of > or =512 + 4 mg/L. When, however, piperacillin/tazobactam was tested as a fixed 8:1 ratio, the MIC distribution became unimodal. If clavulanate 4 mg/L was combined with piperacillin, a unimodal MIC distribution was seen for ESBL-producing Klebsiella spp. but a bimodal distribution arose if the clavulanate concentration was reduced to 0.25 mg/L. These data for alternative combinations suggested that the bimodal MIC distribution seen for piperacillin + tazobactam 4 mg/L was a titration effect, not a reflection of some ESBLs being resistant to tazobactam. Even within single strains, as defined by serotype and DNA fingerprints, there was considerable variation in susceptibility to piperacillin + tazobactam 4 mg/L, with some representatives highly susceptible and others highly resistant. Some of the more resistant representatives produced more of their ESBL, or had a greater number of beta-lactamase types, but these associations were not universal. Elevated resistance to piperacillin + tazobactam was not associated with porin change in any ESBL producer examined, but has been found by others.

Several extended-spectrum beta-lactamases (ESBLs) belonging to molecular Class D have been described from Pseudomonas aeruginosa isolates collected in Turkey. Four of these, OXA-11, -14, -16 and -17, are derivatives of OXA-10 beta-lactamase. We tried to select similar mutants in vitro from OXA-10-producing transconjugants of P. aeruginosa, using a multistep method on ceftazidime-containing agars. Forty-four such mutants were obtained; all had increased resistance to ceftriaxone, cefsulodin, cefepime, cefpirome, latamoxef, aztreonam and, especially, ceftazidime whereas MICs of piperacillin, carbenicillin, cefotaxime, cefoperazone and carbapenems were little altered. Genes related to blaOXA-10 were sequenced from five mutants. One mutant enzyme had aspartate instead of glycine at position 157, and corresponded exactly to natural OXA-14 beta-lactamase. Another mutant strain appeared to have both OXA-14 and a new pI 6.2 enzyme, designated OXA-M102, with serine instead of alanine at position 124 and aspartate instead of glycine at position 157. This latter variant resembled natural OXA-16 enzyme, which has threonine at position 124 and aspartate at position 157. The remaining three mutant enzymes differed from any so far found in wild-type isolates. Two had leucine replacing tryptophan at position 154 (this enzyme was named OXA-M101) while the third (OXA-M103) had a pI of 7.6, and had lysine instead of asparagine at position 143. A different mutation at this position was previously found in OXA-11, a wild-type OXA-10 mutant. Thus, some of the ESBL mutants selected (OXA-14 and OXA-M102) correspond exactly or almost exactly to ESBLs found in wild-types, whereas others (OXA-M101 and OXA-M103) were totally new.

Extended-spectrum beta-lactamases (ESBLs) are an increasing cause of resistance to oxyimino-aminothiazolyl cephalosporins, especially in klebsiellae. In a recent survey we detected ESBLs in 220 (23%) of 966 consecutive klebsiellae from 35 intensive care units (ICUs) in southern and western Europe. The present study examined the extent to which this distribution reflected epidemic strain spread, as against the distribution of ESBL genes into unrelated strains. All 220 ESBL producers were subjected to capsular serotyping and pulsed-field gel DNA electrophoresis (PFGE). Beta-Lactamases were typed for strains isolated on three or more occasions, with the emphasis on SHV enzymes, as these were commoner than TEM variants. Serotyping and PFGE typing defined 85 distinct strains, from 23 of the 35 participating centres. Of 14 centres that contributed five or more ESBL producers, all sent representatives of more than one strain, and two centres sent members of ten or more different strains in contributions of 17-21 ESBL-producing isolates. Nevertheless, epidemic strains-defined as those represented by three or more isolates-accounted for a majority (61%) of the collection. Fifty-two isolates of the same serotype K25 (occasionally acapsular) strain with SHV-4 beta-lactamase were recovered at two French hospitals and one in Belgium. This strain has been found by others in France, and has become particularly widespread. Another single strain was found in two separate Portuguese centres, and many individual hospitals had one or more epidemic strain(s), as well as a scatter of diverse ESBL producers. Major variation in antibiogram and plasmid profile was apparent within strains, with some intra-strain variation in beta-lactamase type. These data imply a fluid situation, with resistance determinants being gained, modified or lost. The endemicity of ESBL producers is disturbing since it limits the potential for control by blocking strain spread, while the diversity within strains is disturbing because it complicates the design of antibiotic policies even during 'single strain' outbreaks.

To test the widespread view that resistance disappears in the absence of antimicrobial use, we tested streptomycin against 477 Enterobacteriaceae from the Royal London Hospital. Twenty per cent proved resistant although streptomycin is little used at the hospital and streptomycin resistance in gram-negative bacteria is caused by mechanisms that do not compromise the drugs that are used. Up to 70% of the observed resistance was associated with cross-resistance to spectinomycin and the presence of ant(3")-Ia, an integron-associated gene carried in Tn21-type transposons. This genetic organization may have conserved streptomycin resistance in the absence of direct selection pressure.

Nineteen isolates of Klebsiella oxytoca were examined, representing 18 distinct strains. All were from a 1994 survey of resistance amongst klebsiellae in intensive care units in Europe, and all had reduced susceptibility, or were resistant, to cefuroxime, ceftriaxone and aztreonam, suggesting hyperproduction of the chromosomal K1 beta-lactamase. We sought to confirm this mechanism and to identify why the levels of resistance varied between isolates. Possible reasons for variation were differences in the quantity or subtype of the K1 enzyme or differences in this enzyme's interplay with permeability. Spectrophotometric assays showed that all 19 isolates had K1-like beta-lactamases and that these were present at > or = 15-fold higher levels than in beta-lactam-sensitive K. oxytoca isolates. Fourteen of the 19 isolates had the OXY-2 form of K1 enzyme, while the remaining five had the OXY-1 form, as determined by isoelectric focusing and PCR amplification. Most isolates with the OXY-2 enzyme were more resistant than those with the OXY-1 subtype, but this difference partly reflected enzyme quantity rather than subtype. More generally, and irrespective of enzyme subtype, levels of resistance were broadly related to beta-lactamase specific activity, and the degree of hyperproduction was a major determinant of the level of resistance. Nevertheless, other factors had a role too: several isolates had reduced susceptibility or were resistant to cefoxitin, which is not a substrate for K1 enzyme, and examination of outer membrane protein profiles revealed considerable strain-to-strain diversity in the molecular weight range typical of the major enterobacterial porins (40-48 kDa).

PER-1, an extended-spectrum class A beta-lactamase, has been described only from Pseudomonas aeruginosa RNL-1, which was obtained in France in 1991. During studies on ceftazidime-resistant P. aeruginosa collected from December 1991 to July 1993 in Ankara, Turkey, we found 14 further isolates with PER-1 enzyme, as recognised by isoelectric point (pI 5.4), hydrolytic activity, gene hybridization and DNA sequence. Five of these isolates also had OXA-10 (PSE-2)-related enzymes and one hyper-produced ampC beta-lactamase, whereas eight had PER-1 enzyme alone. The last group, from four wards, appeared to be identical, giving the same DNA restriction patterns and carrying the PER-1 gene on an 8.5 kb HincII fragment. Two more producers were related but the other four were unique. In several representatives of the group of eight replicates, the PER-1 gene was shown to be encoded on a plasmid, larger than 154 kb in size, which transferred to P. aeruginosa PU21. A further isolate had the gene on an even larger conjugative plasmid. By contrast, the PER-1 gene reportedly was chromosomally-inserted in strain RNL-1. The PER-1 producers and their transconjugants were highly resistant to ceftazidime and aztreonam (MIC > or = 128 mg/L) but not to carbapenems or latamoxet. Piperacillin insusceptibility was marginal (MIC 8 mg/L). Clavulanate 4 mg/L, but not tazobactam 4 mg/L, reversed resistance to ceftazidime and carbenicillin. Purification of the enzyme to homogeneity was achieved by three ion exchanges and one gel filtration. We found much lower Vmax rates for aminothiazolyl cephalosporins than reported previously for PER-1 enzyme. This reflected the present assays being in 0.1 M phosphate buffer pH 7.0, whereas the previous were pH-stat-regulated; concentrated phosphate reduced enzyme activity against ceftazidime, but not against cephaloridine.

beta-Lactamase expression was examined in 1000 consecutive Gram-negative bacilli isolated from urine, wound swab, sputum or blood specimens received at the Microbiology Laboratory of the Royal London Hospital. This survey, performed between January and April, 1991, followed a similar study undertaken in early 1982. The distribution of species was similar in the two surveys, except that the proportion of Pseudomonas aeruginosa isolates had increased from 11% in 1982 to 17.5% in the present study. This increase was balanced by a decreased proportion of enterobacteria. Amongst plasmid-mediated beta-lactamases, TEM-1 (especially), TEM-2, SHV-1 and OXA types continued to predominate in enterobacteria. Their frequency in Escherichia coli was unchanged (46% in 1991 compared with 43% in 1982), but had increased from 5 to 22% amongst Proteus mirabilis isolates. An apparent decrease in their frequency amongst Enterobacter cloacae isolates, from 48% in 1982 to 17% in 1991, probably reflected changes to strain prevalence rather than enzyme prevalence. Plasmid type beta-lactamases were present in fewer than 2% of P. aeruginosa isolates in both surveys. In the present study, chromosomal beta-lactamase derepression (constitutive hyperproduction) was detected in 10/76 isolates of E. cloacae, Enterobacter aerogenes, Citrobacter freundii, Serratia spp. and Morganella morganii, and in 2/170 P. aeruginosa isolates. These proportions were increased, compared with those seen the 1982 survey, though the significance was borderline (P approximately 0.05; chi 2 test). Extended-spectrum plasmid mediated beta-lactamases, unknown in 1982, were found in 11/70 Klebsiellae pneumoniae isolates in the present study. Ten of these organisms, representing at least five distinct strains, produced TEM-10 enzyme, encoded by a plasmid of c. 90 kb; the remaining organism had an unidentified SHV-derived enzyme.

Recent reports of decreased susceptibility to amoxycillin/clavulanate and ticarcillin/clavulanate combinations amongst Escherichia coli isolates have been attributed to an increased frequency of hyperproduction of TEM-1 beta-lactamase. To test this claim we compared the activities of clavulanate combinations against consecutive E. coli isolates producing TEM-1 beta-lactamase that were obtained from clinical material at The London Hospital during 1982 (n = 50) and 1989 (n = 46). Enzyme production was quantified and related to the level of clavulanate required to potentiate amoxycillin and ticarcillin. The levels of TEM-1 enzyme production varied 150-fold amongst the isolates, partly because of variation in the gene copy number. A clear correlation existed between enzyme quantity and levels of resistance to the clavulanate combinations. However, there was no significant difference (P greater than 0.05, Mann-Whitney U test) between the isolates obtained in 1982 and 1989 in terms of the clavulanate concentrations required to potentiate the penicillins, or in the distribution of beta-lactamase activities present.

A 30 base synthetic oligonucleotide probe to the AAC(6') coding region of the gentamicin resistance transposon Tn4001 was constructed. This was used to probe plasmid extracts and restriction endonuclease (RE) digests of total DNA from epidemic methicillin-resistant Staphylococcus aureus (EMRSA) isolates. Hybridization occurred to plasmid bands of gentamicin-resistant EMRSA which carry a gentamicin resistance plasmid, but not to DNA of gentamicin-sensitive EMRSA. Hybridization also occurred to RE digestion fragments of total DNA from recently-isolated, plasmid-free gentamicin-resistant EMRSA isolates. This confirmed the presence of the gentamicin-resistance gene in the chromosomal DNA of the 'new' EMRSA isolates.