Tools for microbial detection and characterization in drinking water distribution systems

Safe drinking water is essential for public health. Recent outbreaks in industrialised countries have highlighted human vulnerability to pathogens found in drinking water supplies. Effective monitoring of water supplies is essential in order to prevent waterborne disease. Total coliforms and Escherichia coli (E. coli) are measured as indicators of potential faecal contamination of the water supply. Any method used by water utilities to detect and identify a microbial isolate from the drinking water supply must be reliable with the ability to provide results with a high degree of sensitivity (true positives) and specificity (true negatives).Many rapid tests for coliforms rely on the ability of coliforms to utilize O-nitro-beta-D-galactopyranoside (ONPG). Rapid tests for E. coli rely on this bacterium's ability to degrade both ONPG and 4-methylumbelliferyl-beta-glucuronidide (MUG). In order to utilize ONPG the bacterium must have the enzyme beta-D-galactosidase, which is encoded for on the lacZ gene. In order to utilize MUG, the bacterium must have the enzyme beta-D-glucuronidase, which is encoded for on the uidA gene. Previous research has shown that some coliform isolates and E. coli isolates in the presence of chlorine have been O-nitro-beta-D-galactopyranoside (ONPG) activity negative, which results in false negatives. Likewise there are environmental and pathogenic strains of E. coli including E. coli O157:H7 which do not metabolise MUG and therefore, would result in false negative results for E. coli detection.The intention of this research is to expand the understanding of false negative coliform and E. coli occurrences in drinking water distribution systems. Through an examination of disparities between methods, and an analysis of enzyme utilization of bacteria isolated during disinfection it may be possible to create a more thorough understanding of the limitations of the current detection methods. A multi-tier framework that incorporates traditional membrane filtration, defined substrate testing and QPCR as well as enzyme utilization characterization was developed to compare current relevant identification and enumeration methods under various disinfection scenarios.The experiments used Annular Reactors (ARs) equipped with polycarbonate coupons to simulate a drinking water distribution system. The ARs were inoculated with E. coli or coliforms. There was also a control AR that received no inoculum. The ARs were disinfected using chlorine residuals in the range of 0.2-0.3 mg/L. Effluent water and biofilm from polycarbonate coupons were sampled and analysed for coliforms and E. coli using membrane filtration on mEndo agar, two defined substrate methods (Colilert and Colitag) and QPCR. Isolates were subsequently characterized on API 20E test strips to analyse for enzyme usage profiles. Isolates were also analysed using QPCR to test for the presence of the lacZ and uidA gene.There are a variety of enzyme based and molecular based tools available for the analysis of drinking water including traditional membrane filtration onto selective agar such as mEndo, defined substrate testing using chromogenic and fluorogenic indicators for coliforms and E. coli respectively, and molecular methods such as quantitative polymerase chain reaction (QPCR) which can detect the presence of the lacZ and uidA gene sequences found in coliforms and E. coli, respectively. However, each of these detection methods has limitations and is vulnerable to presenting false negative and false positive results.Although the ONPG negative phenotypes were not observed, it was shown that there was a significant discord between culture methods and defined substrate methods. The defined substrate method Colilert did not identify some samples as E. coli due to lack of MUG activity, although all the isolates taken from the same bulk sample were shown to be E. coli when analysed by biochemical substrate test strips (API 20E) and QPCR. The framework approach allowed for a more thorough investigation of deviations between detection methods, and helped to offer some explanations for those deviations including changes in enzyme profiles and the presence of indicator DNA sequences.The framework used in this thesis work was able to successfully analyse drinking water in the laboratory. By including improvements to the QPCR method and expanding the characterization and confirmation steps to directly analyse the defined substrate methods, the framework could become a valuable tool for the analysis of water in a distribution system.