Ability of three DNA-based assays to identify presumptive Escherichia coli colonies isolated from water by the culture-based mFC agar method

We tested the ability of three PCR assays, targeting uidA and tuf genes to correctly identify Escherichia coli colonies isolated from water and we compared them to two β-glucuronidase-based culture methods (Colilert ® and Readycult ®), in terms of specificity and sensitivity. E. coli isolates recovered on mFC agar were first tested for the presence of the uidA positive colonies were presumed to be E. coli. For further characterization, uidA-negative colonies were subsequently identified using the Vitek 2 automated system. Colilert ® and Readycult ® detected 436 and 442 of 468 colonies identified as E. coli on mFC corresponding to sensitivities of 93.2 and 94.4%, respectively. None of the 59 non- E. coli isolates was detected by both methods for a specificity of 100%. Two (2) uidA and 1 tuf PCR assays were also tested. The uidA PCR assays yielded positive signals for 447 (95.5%) and 434 (92.7%) of 468 E. coli isolates tested respectively, whereas the tuf PCR assay showed a sensitivity of 100%. None of the 59 non- E. coli isolates was detected by both uidA PCR assays (100% specificity), whereas tuf PCR false-positive signals were obtained with Escherichia fergusonii and Escherichia albertii. However, since these 2 species are principally found in the feces of mammals and birds, their detection indicates a fecal contamination. Consequently, using a 1-h tuf rtPCR assay to confirm the identity of E. coli colonies on mFC agar is as specific, more sensitive, and potentially more cost-efficient than culture methods based on β-glucuronidase detection. ► mFC agar detects 5–8% more E. coli in water than culture methods based on β-glucuronidase activity ► uidA PCR was 100% specific but failed to detect 4.5–7.3% of E. coli ► tuf rtPCR identified 100% of E. coli but detected E. fergusonii and E. albertii