RESUMEN
Underestimation of Escherichia coli in drinking water, an indicator microorganism of sanitary risk, may result in potential risks of waterborne diseases. However, the detection of disinfectant-injured or genetically modified (GM) E. coli has been largely overlooked so far. To evaluate the accuracy of culture-dependent enumeration with regard to disinfectant-injured and GM E. coli, chlorine- or ozone-injured wild-type (WT) and GM E. coli were prepared and characterized. Then, water samples contaminated with these E. coli strains were assayed by four widely used methods, including lactose tryptose broth-based multiple-tube fermentation (MTF), m-endo-based membrane filtration method (MFM), an enzyme substrate test (EST) known as Colilert, and Petrifilm-based testing slip method (TSM). It was found that MTF was the most effective method to detect disinfectant-injured WT E. coli (with 76.9% trials detecting all these bacteria), while this method could not effectively detect GM E. coli (with uninjured bacteria undetectable and a maximal detection rate of 21.5% for the injured). The EST was the only method which enabled considerable enumeration of uninjured GM E. coli, with a detection rate of over 93%. However, the detection rate declined to lower than 45.4% once the GM E. coli was injured by disinfectants. The MFM was invalid for both disinfectant-injured and GM E. coli. This is the first study to report the failure of these commonly used enumeration methods to simultaneously detect disinfectant-injured and GM E. coli. Thus, it highlights the urgent requirement for the development of a more accurate and versatile enumeration method which allows the detection of disinfectant-injured and GM E. coli on the assessment of microbial quality of drinking water.