Background Aminoglycosides are one of the clinically relevant antibiotics. Plasmid-encoded 16S rRNA methylase enzymes have emerged in clinical isolates of Gram-negative bacteria worldwide. The spread of these resistance determinants has become a great concern. Objectives The objectives of this study were to investigate the prevalence of 16S rRNA methylases and aminoglycoside modifying enzyme, AAC(3)-II in Escherichia coli isolated from human blood cultures and faecal samples of animals. E. coli isolates with unexplained aminoglycoside resistance phenotypes were investigated by detection of four aminoglycoside modifying enzymes, AAC(6’)-I, ANT(2”)-I, ANT(4’)-II and APH(3’)-VI. Methodology This study included 188 E. coli clinical isolates obtained from blood cultures of patients in one regional hospital between January 2004 and September 2010 and 81 E. coli isolates obtained from faecal samples of chickens, pigs, cattle, cats, dogs and rats between September 2008 and August 2011. All 269 E. coli isolates in this study were screened for the aac(3)-II gene and six 16S rRNA methylase genes(armA, rmtA, rmtB, rmtC, rmtD and rmtE)by two individual sets of multiplex PCR assays. A subset of 88E. coli isolates with aminoglycoside resistance phenotypes, which could not be explained by the genes detected, were subjected to detection of the aac(6’)-Ib, ant(2”)-Ia, ant(4’)-IIaand aph(3’)-Via genes by four individual PCR assays. The transfer of resistance of the rmtB gene was studied by conjugation experiments. The clonal relationship between rmtB-producing strains was investigated by pulsed-field gel electrophoresis. Results 67.6% (25/37) and 63.4% (26/41) of the Gen-R/Amk-NS group isolates from human and animal sources, respectively, were found to possess the aac(3)-IIgene. The aac(3)-IIgene was also found in 96.7% (146/151) Gen-R/Amk-S group human isolates. 21.6% (8/37) and 61%(25/41) of the Gen-R/Amk-NS isolates from human and animal sources, respectively, were found to possess the rmtB gene. The armA gene was found in one human and one animal isolates, which were both resistant to gentamicin and amikacin. No rmtA, rmtC, rmtD orrmtE genes were found in this study. Among 88E. coli isolates with unexplained aminoglycoside resistance phenotypes, the aac(6’)-Ib gene was found in51.2%(22/43) and 10% (4/40) of the Gen-R/Amk-NS group and the Gen-S/Amk-NS group, respectively. The ant(2”)-Ia gene was found in 11.6% (5/43) of the Gen-R/Amk-NS group E. coli isolates. No ant(4’)-IIa or aph(3’)-Via genes were found. No major PFGE cluster was observed among 32 rmtB-positive isolates by pulsed-field gel electrophoresis.In addition, amikacin resistance could be transferred by conjugation from 12rmtB-positive donors. Conclusion The present study showed that the rmtB gene was the most prevalent 16S rRNA methylase gene in both human and animal E. coli isolates. A high incidence of the aac(3)-IIgene was found among gentamicin-resistant strains. The spread of 16S rRNA methylases has aroused clinical concern and become a major therapeutic threat in the future. / published_or_final_version / Microbiology / Master / Master of Medical Sciences

published_or_final_version / Microbiology / Doctoral / Doctor of Philosophy

Dissemination of CTX-M type extended-spectrum beta-lactamase (ESBL)-producing Escherichia coli (E. coli) is a serious health issue in Hong Kong. However, research knowledge concerning its dissemination mechanism and plasmid characteristics over time in health care setting is still limited. This study was conducted to characterize ESBL-producing E. coli from blood culture isolates and the epidemiology of the plasmids harboring CTX-M-14 collected from a healthcare region in Hong Kong from two periods of time. A total of 167 ESBL-producing E. coli in blood culture specimens were retrieved from period 1 (1996 - 1999, n = 50) and period 2 (2007 - 2008, n = 117). Antimicrobial susceptibilities were defined by disk diffusion method according to CLSI. Phylogenetic groups and CTX-M enzymes were detected among all the ESBL-producers. Clonal relatedness of the hosts was analyzed by pulsed-field gel electrophoresis and multi-locus sequence typing. A subset of 65 CTX-M-14-producing isolates was undergone for further plasmid characterization. Conjugation, PCR-based replicon typing, S1-PFGE, southern-blot hybridization, and genetic environment PCRs were performed. Plasmid PCR-restriction fragment length polymorphism (pRFLP), F-allele replicon sequence typing and variable region PCRs were studied in 54 F-plasmids obtained. Results showed that over half of the ESBL-positive isolates were non-susceptible to ciprofloxacin, cotrimoxazole and gentamicin. A surprisingly high number of CTX-M-type ESBL was carried by 98.2% (164/167) of the isolates. CTX-M-9 group (89.8%, 150/167) and CTX-M-14 (103/109) were predominantly found among both periods. Overall, nearly half (41.3%, 69/167) of the isolates belonged to 5 major clones. Clonal types undetermined-ST68 (n = 18) and O102-ST405 (n = 15) were dominant in period 1 while clonal types O25b-ST131 (n = 30), O15-ST69 (n = 5) and O12-ST12 (n = 3) emerged in period 2. Among a subset of 65 CTX-M-14 plasmids, most of them were transferable (84.6%, 55/65) with high frequency, similar plasmid sizes and genetic environment ISEcp1-blaCTX-M-14-IS903 (90.8%, 59/65). Replicon types of the CTX-M-14 encoding plasmids were FII (n = 48) or FII ± FIA/FIB types (n = 6), I1-I (n = 3), B/O (n = 2), K (n = 1) and undetermined (n = 4). Subtyping of 54 IncF plasmids by replicon sequence typing, pRFLP and PCR for marker genes (yac, malB, eitA, eitC and parAB) showed that 79.6% (43/54) of the plasmid subset exhibited identical or highly similar results with the completely sequenced plasmid, pHK01 (E. coli isolated from urine sample of a patient in Hong Kong, 2004). These 43 plasmids were originated from both period 1 (n = 11) and period 2 (n = 32). These pHK01-like plasmids were found to have spread to the major clones (ST68, ST405 and ST131) and multiple singleton isolates of all four phylogenetic groups. In conclusion, this study demonstrated the widespread dissemination of pHK01-like CTX-M-14 encoding plasmids among isolates of diverse genetic lineages over a decade. The dissemination was probably due to both clonal expansion and horizontal gene transfer of pHK01-like IncF plasmid. / published_or_final_version / Microbiology / Master / Master of Philosophy

Not available / text

published_or_final_version / abstract / Microbiology / Master / Master of Philosophy

published_or_final_version / Botany / Master / Master of Philosophy

Early events occurring during induction of the SOS response in Escherichia coli are poorly understood. In order to understand the early steps in SOS induction more fully, we have isolated several mutations which constitutively express the SOS regulon. Using a Mud(Apᴿ,lac) fusion to the SOS regulated gene sulA, we isolated Lac⁺ colonies as mutants in which RecA protein is constitutively activated for repressor cleavage. The mutations map to four loci: dam, lig, uvrD and recA. The extent of constitutive SOS induction in these mutants varied greatly, indicating different levels or types of signal in the cell. The mutations isolated demonstrate two early steps in SOS induction. The first step in SOS induction is signal generation and includes mutations found in dam, lig and uvrD genes. The mutant gene products presumably alter DNA metabolism to produce an inducing signal. These non-lethal mutations lead to sub-induction and probably generate very specific signals, such as abnormally unwound DNA in the case of DNA helicase II mutants or unsealed DNA nicks that result from deficient ligation in lig mutants. Greater induction may require quantitatively more signal or different types of signal generated by severe defects leading to cell death. These mutations also show that signal is a variable quantity, allowing the cell to fine tune the levels of SOS repair activity according to the amount or type of signal (damage) perceived. In some cases (such as dam mutations), blocking the SOS response by lexA(Ind⁻) alleles leads to cell death. In this type of constitutively activated strain, the increased level of repair from SOS induction is required to allow the cell to tolerate potentially lethal DNA structures generated by the mutant gene product. The second step in induction is the interaction of signal with RecA protein and is shown by isolating 8 recA mutants. Mutant recA alleles caused the strongest SOS induction in any mutants obtained, similar to the level found in strains lacking repressor (lexA(Def) mutants). This full induction in the absence of lethal DNA damage underscores the pivotal role of RecA protein in regulating the SOS response.

published_or_final_version / abstract / Microbiology / Master / Master of Philosophy

Previous genetic evidence suggested that RecA was required in SOS mutagenesis for its regulatory role and perhaps some other nonregulatory role (Mount, 1977; Blanco et al., 1982). I undertook a genetic study which confirmed the above studies and provided further evidence that RecA protein appeared to have a dual "role in mutagenesis; first, the cleavage of LexA repressor for the derepression of specific SOS genes and second, one or more additional role(s). For these studies a new phage mutagenesis assay was developed which allows rapid scoring of SOS mutagenesis in a large number of host mutants. I next conducted a genetic analysis to determine if the newly defined RecA mutagenesis function was separable by mutation from the numerous other phenotypes which are known to be influenced by RecA protein. From the study of recA mutants it appears that the RecA mutagenesis function(s) is genetically separable from the following RecA phenotypes: LexA cleavage, lambda cI repressor cleavage, UV resistance and homologous recombination. In addition, I discovered that the LexA cleavage function and lambda cI cleavage function is also separable. I also studied in some detail the novel genetic properties that I uncovered for recA432 mutant strains. recA432 was defined as a mutagenesis defective allele (Kato and Shinoura, 1977). LexA cleavage in recA432 cells was more easily induced that in recA⁺ cells, causing lethal filamentation of these mutant cells even at very low UV doses. I concluded that the basis for the Mut⁻ phenotype was this strain's propensity to lethally filament, which complicated the detection of mutant cells. In another set of experiments, I examined the regulatory requirements for SOS mutagenesis and Weigle phage-reactivation; I wanted to determine which SOS operons must be derepressed for this process. lexA(Ind⁻) mutant cells are defective in mutagenesis because they cannot derepress specific SOS genes required in this process. I found that the selective derepression of umuDC was sufficient to restore mutagenesis to these lexA(Ind⁻) mutants; however, derepression of umuDC and recA was required for phage reactivation.

The LexA protein of E. coli is a repressor of at least 20 genes in the SOS regulon, and by this function plays a major role in regulating the SOS response. Two different genetic approaches have been taken to define the DNA binding domain of LexA repressor. First, several mutant repressors which are defective in DNA binding have been isolated. The mutations generating these repressors were dominant to lexA+, indicating that the mutant proteins can act in trans to interfere with binding of normal repressor to an operator sequence. The repressors may be defective due to elimination or disruption of contacts made between side chain(s) within the protein and the DNA helix but dominant because they can still interact with other monomers of LexA protein. In a second experiment to define the DNA binding domain of LexA protein, a novel genetic selection has been used to isolate DNA binding specificity mutants. The recA operator (CTG TATGA.GCATA CAG), a known lexA binding site, has been altered in a symmetric fashion. This choice was based on the assumption that the dyad symmetry of the operator indicates at least two repressor monomers bind to each operator such that each monomer recognizes one half of the operator. A class of mutant repressors which restored binding to this altered operator but had little affinity for the wild-type recA operator was isolated. This type of mutation allowed the identification of amino acids in the repressor which are likely to make specific contacts with base-pair(s) in the DNA binding site. By examining the effects of a series of amino acid substitutions on repressor specificity, it was possible to show that a glutamic acid residue at position 45 (E45) contacts the first and last base-pair of the consensus recA operator (CTG TATGA.GCATA CAG). Both negative dominant and operator recognition mutations were located in a small region that was previously identified to specify a helix-turn-helix motif based on sequence similarity to other repressors. These studies therefore suggest that LexA protein may bind to DNA by a helix-turn-helix motif similar to these repressors.