Loss of Antibiotic Resistance: Analysis of Phenotype and Related Gene Expression

Progress: Initial work involved collection of libraries of Campylobacter and Salmonella and then analysing these strains for resistance to antibiotics, disinfectants, dyes and organic solvents and for presence of resistance genes. For E. coli and Salmonella resistance to certain antibiotics, disinfectants and dyes and resistance to organic solvents has been attributed to up-regulated efflux via the efflux pump AcrAB and to up-regulation of global regulators such as marRAB and soxRS. <P>
For Salmonella there was a clear link between resistance to cyclohexane and low level resistance to MAR-type antibiotics, disinfectants and dyes. Cyclohexane resistance was seen in ten different serotypes (mainly from poultry) and comprised 10.6% of the panel tested. By comparison with earlier data (strains selected pre-1996) there has been a < 5% increase in cyclohexane resistance. <P>
Of 65 cyclohexane resistant isolates 49 were examined further. These comprised 21 cyclohexane resistant isolates of which 13 were ciprofloxacin resistant isolates, 29 cyclohexane sensitive but ciprofloxacin resistant strains with MIC >0.25µg/ml ciprofloxacin, and 9 strains which were susceptible to both cyclohexane and ciprofloxacin and served as controls. The 47 isolates comprised 15 different serovars. Expression levels of the two global regulators, marA and soxS, and the efflux pump gene acrB, were measured and compared to wild type expression levels to establish changes in expression. Further experiments to confirm the data and perform detailed statistical analyses are in progress. <P>
For Campylobacter there was no clear link between resistance to various organic solvents tested and low level antibiotic resistance, however, detailed analysis of strains showed some link between reduced susceptibility to acridine orange, ethidium bromide and triclosan and reduced susceptibility to some antibiotics. Such strains showing reduced susceptibility (or a MAR-like phenotype) were further analysed. Seven MAR isolates over-expressed cmeB, 3/32 over-expressed cmeB and cmeF, and none over-expressed cmeF alone. Expression of porA was similar in all isolates. In conclusion, over-expression of cmeB occurred more frequently than that of cmeF. These data suggest that CmeB may be associated with the MAR phenotype. However, although CmeB has previously been shown to transport a variety of agents, it was over-expressed by only 31.3% of the MAR-like isolates, suggesting other mechanism(s) of a MAR type in C. jejuni. The Campylobacter strains were also tested for the presence of the enzyme ?. lactamase and for the resistance genes tet(0) and tet(M). <P>
The panel of Salmonella strains was also tested for the presence of the antibiotic resistance genes aadB, aphAI-IAB, aadA1, aadA2, bla(Carb2) or pse1, bla(Tem), cat1, cat2, dhfr1, floR, strA, sul1, sul2, tetA(A), tetA(B) and tetA(G) genes and for the presence of class 1 integrons . The aim of this was to clearly identify genes involved in higher level non-MAR resistance, to determine the epidemiology of these resistance genes, their linkage with integrons and the level of resistance conferred by different genes. Class 1 integrons were found in S. Emek, Haifa, Heidelberg, Mbandaka, Newport, Ohio, Stanley, Virchow and in Typhimurium, mainly phage types DT104 and U302 and these strains were generally multi-resistant up to seven antibiotics. Certain genes such as bla(Carb2), floR and tetA(G) were almost exclusively associated with S. Typhimurium DT104 and this correlates with previous findings, although there have been rare exceptions. Other antibiotic resistance genes such as aadA2, strA, sul1 and sul2 were associated with a wider range of serotypes whilst others such as aadA1 were only found in non-Typhimurium isolates. Where present antibiotic resistance genes were expressed with the exception of some strains harbouring the strA gene which were sensitive to streptomycin. <P>
For Salmonella the mutation rate in the presence of MAR-antibiotics was determined for strains with and without prior disinfectant passage. Resulting mutants were analysed for the MAR phenotype (resistance to non-related antibiotics and cyclohexane resistance). Prior passage with disinfectant, particularly the disinfectant triclosan, led to a statistically higher occurrence of MAR-like mutants when strains were plated onto agar with antibiotics. Resulting mutants are being analysed for mutations in the acrOR and marOR regions. Passage in the presence of non-MAR antibiotics led to antibiotic resistant mutants, but these were not shown to have the MAR phenotype evidenced by sensitivity to cyclohexane. E. coli exposed to ciprofloxacin was also shown to develop the MAR phenotype and micro-array studies showed that two adaptive mutants over expressed the marRAB [2].
For C. jejuni NCTC11168 passage with cefotaxime led to an 8 to 32 times increase in resistance to the un-related antibiotics ampicillin, chloramphenicol, ciprofloxacin, erythromycin and tetracycline and this was associated with over-expression of cmeB. Passage of this strain with just ciprofloxacin led only to increased ciprofloxacin resistance associated with mutation in gyrA. However, two derivative efflux pump knockout strains, cmeB::kanR and cmeF::kanR when exposed to ciprofloxacin gave rise to some mutants with resistance to ciprofloxacin, chloramphenicol, tetracycline, ethidium bromide, acridine orange and SDS. These strains had no mutation in gyrA or altered expression of the efflux pump genes cmeB, cmeF or the major outer membrane protein gene porA.
In conclusion, this work has successfully analysed a wide range of field strains of Campylobacter and Salmonella for the MAR phenotype and investigated the genes involved in this low-level resistance as well as the genes involved in higher level resistance. Work has also been carried out to look at the development of the MAR phenotype and it was shown that for both Campylobacter and Salmonella passage with antibiotics and / or disinfectants gave rise to the MAR phenotype and genes involved have been investigated. For Salmonella data suggested that exposure to disinfectant and then selection of low level mutants with antibiotic may be an important double selection procedure for MAR mutants in the field. The data also suggested that for Salmonella the MAR phenotype may be becoming more common amongst different serotypes of Salmonella. It is possible that that changes in on-farm, food treatment or human health-care practices has increased the selective pressure for the acquisition of low level resistance and this is an area for future vigilance. <P>
In these studies, with the exception of 8/59 strains which were positive for the strA gene but were sensitive to streptomycin, there was no evidence for the presence of resistances gene without the appropriate phenotype suggesting that the loss of a resistance phenotype and retention of a silent gene might be a very rare event. However, this study showed the converse, namely that strains could readily acquire low level of multiple antibiotic resistance by mutation in gene encoding and regulating efflux systems without the presence of a classic resistance gene.

<P>