Aptness of Escherichia coli host strain CB390 to detect total coliphages in Colombia

Fecal bacteria have been used for more than a century as indicators of fecal contamination in water. In recent years, the monitoring of somatic and F-specific coliphages has been gradually included in guidelines and regulations as an additional parameter to reinforce water safety. The Escherichia coli host strain CB390 was tailored to detect both somatic and F-specific coliphages in a single test. The efficacy of this strain for bacteriophage detection, previously evaluated in Western Europe and North America, was assessed here for the first time in South America. The detection of somatic and F-specific coliphages by the strain CB390, as well as by standardized methods, was performed in drinking and river water and municipal and abattoir wastewaters. No statistical difference was found in the numbers of total coliphages detected by strain CB390 and the sum of somatic and F-specific coliphages determined separately by the standardized ISO methods. The data presented here provide further validation of the effectiveness of the host strain E. coli CB390 for the detection of total coliphages in waters in a single test and demonstrate its suitability for application in upper-middle income countries of the Americas (World Bank category).

www.nature.com/scientificreports www.nature.com/scientificreports/ Thus, concentrations of E. coli, and somatic and F-specific coliphages detected by the E. coli strain CB390 were determined in an array of samples consisting of urban and abattoir wastewaters, river water and tap water. Considering the worldwide applicability of WHO guidelines for potable water reuse (13), the data provided here is relevant and useful.

Materials and Methods
Samples. One hundred and ninety-eight water samples with different levels of fecal contamination, representing diverse settings, were collected and analyzed over a 2-year period. Somatic and F-specific coliphages, bacteriophages infecting E. coli strain CB390 and E. coli, were analyzed from 48 samples of urban sewage, 84 samples of the Bogotá River collected in different sectors of the river with very different fecal contaminant loads, and 46 tap water samples of the Bogotá network. Regarding sewage samples, 18 of them were influent sewage of wastewater treatment plants and 30 were secondary effluents. Samples of drinking water were collected from faucets inside private homes throughout a marginal neighborhood, Ciudad Bolívar, occupied by citizens displaced by the political violence of the last decades. Inhabitants from this district normally use small homemade tanks for water storage and take water from taps that are fitted with gadgets to prevent water splashing. In addition, E. coli, somatic coliphages and bacteriophages infecting strain CB390 were analyzed in 20 samples of raw wastewater of abattoirs slaughtering porcine and bovine specimens.
Sample processing. Before the phage analysis, wastewater, secondary effluents and river water samples were decontaminated by filtration through 0.22 µm diameter pore-size Millex-GP membranes, which are low-protein-binding polyethersulfone membranes (Millipore Corporation, Bedford, MA, USA).
The bacteriophages from samples corresponding to a section of the river with low fecal contamination and drinking water from the distribution network were concentrated from 1 liter samples in accordance with Méndez et al. 30 . Briefly, phages in 1000 mL of water samples were concentrated by adsorption to 0.22 µm pore-size cellulose ester membrane filters (GSWP; Millipore Corp., Bedford, MA), followed by elution with a solution of 1% beef extract, 0.05 M NaCl, and 3% Tween 80, and ultra-sonication.
Bacteria enumeration. E. coli were enumerated by membrane filtration using Chromocult coliform agar (Merck, Darmstadt, Germany) supplemented with antibiotics (E. coli/coliform selective supplement; Merck). Dark-blue/purple colonies were presumed to be E. coli and confirmed by the addition of Kovac's reagent 31 . Bacteriophage enumeration. Somatic coliphages were quantified by enumerating plaque forming units (PFUs) in the E. coli host strain WG5 according to the ISO 10705-2 standard method 18 . F-specific coliphages were quantified by enumerating PFUs in strain WG49 of Salmonella enterica serovar typhimurium according to the ISO 10705-1 standard method 17 . To quantify somatic and F-specific coliphages in a single assay, host strain CB390 was grown in MSB broth with 100 µg per ml of ampicillin. TYG agar and TGY semisolid agar 17 supplemented with www.nature.com/scientificreports www.nature.com/scientificreports/ ampicillin (100 g ml −1 ), Ca 2+ , and Mg 2+ , as in the ISO 10705-2 18 standard procedure, were used for the lower and upper layers in the double-agar-layer test, as indicated in Guzmán et al. 27 .
Data treatment and statistics. The sum of the counts of somatic and F-specific coliphages was considered as total coliphages. To avoid dispersion due to the very different levels of fecal pollution of the tested waters, especially the river samples, the box-plots were calculated associating samples according to E. coli concentration intervals rather than by water origin.
Statistical analysis was carried out using R software package version 3.5.1 32 and RStudio version 1.465 33 . Raw data were plotted as box-plots; the calculations performed to generate the box-plot graphs included quartiles and outlier values for each group of samples. Additionally, the data were plotted as quantile-quantile plots (qq-plots). Outlier values were determined according to Tukey's method and were omitted from the statistical comparison tests.
Since not all data were normally distributed, the Wilcoxon rank sum test with continuity correction, the Kolmogorov-Smirnoff test and a two-sample Anderson-Darling test 34 were carried out to compare the series of data. Whenever it is mentioned that differences were statistically significant or not, this is true for all three tests.
Values of coliphages and E. coli concentrations in the drinking water, where the positive detection frequency was very low, were estimated by applying Thomas' formula for the calculation of the most probable number for long series of data 35 . Figures 1 and 2 show the counts of the different indicators studied in urban and abattoir wastewaters and surface waters, indicating magnitude, order of abundance and the proportions among them, as described in the related literature 21 . These results are of interest, since there is a lack of comparative data for total coliphages detected by strain CB390 or other coliphages in the Americas.

Results and Discussion
The values obtained for drinking water are shown in Table 1. Worthy of mention, although out of the main scope of the current work, are the low values of fecal contamination found in drinking water samples, which is a positive result for a marginal neighborhood like the one studied here. The predominance of E. coli over the coliphages points to a source of contamination in the neighborhood, perhaps in the depots or due to cross-contamination, since water disinfection leads to higher proportions of coliphages than E. coli 36 . Unfortunately, the values of coliphages were too low to derive significant conclusions regarding the aptness of strain CB390 for counting total coliphages.  www.nature.com/scientificreports www.nature.com/scientificreports/ No statistical difference (p > 0.05) could be appreciated when the values of somatic coliphages, strain CB390 and the total coliphages were compared. Only the values of F-specific coliphages were significantly lower (p < 0.05) than the others. A comparison of the different bacteriophage counts in the waste and surface waters, sample-by-sample, is visualized in Fig. 2, as in other studies on somatic and F-specific coliphages in similar waters 5,21 . As described elsewhere 25,27,28 , the inclusion of F-specific phages to obtain total coliphages did not result in significantly higher values (p > 0.05) than those of somatic coliphages in the kinds of water tested. Additionally, the counts obtained by strain CB390, though slightly lower, are not significantly different (p > 0.05) from those obtained by the ISO procedure for somatic coliphages and total coliphages. This has also been observed in Europe 25,27 and the USA when using US EPA methods to detect somatic and F-specific coliphages 28,29 . Therefore, these data reinforce the conclusion that the E. coli host strain CB390 is effective for the simultaneous detection of the total number of coliphages in different types of fecally contaminated water. Moreover, it is suitable for application in South American countries with upper-middle incomes.