Development of a qPCR assay for the detection of naturalized wastewater E. coli strains

Sewage contamination of surface water sources used for drinking, recreation, and/or irrigation is a global public health challenge (Leclerc et al. 2002; Garcia-Aljaro et al. 2019). The United States Environmental Protection Agency (U.S. EPA) has estimated that as many as 40,000 sewer overflows occur each year in the U.S., and up to 500,000 km of coastlines, rivers, and streams do not currently meet ambient water quality guidelines due to human and animal waste contamination (U.S. EPA 2007a, 2007b). Sewage contamination can enter environmental waters through several routes: (i) inadequately treated sewage discharged directly into surface water sources; (ii) combined sewer overflows (CSOs); and (iii) leaking septic tanks or sewer lines. As human sewage may impact water bodies used for human recreation, drinking, or food production (i.e., irrigation), determining the source attribution of fecal contamination in these sites is important for evaluating public health risks (Soller et al. 2010).

Several human fecal/sewage-specific indicators have been characterized, including human-specific viruses [e.g., adenovirus, enterovirus, and polyomavirus (Jiang et al. 2001; van der Sanden et al. 2013; Rusinol et al. 2016)], protozoan parasites [e.g., Cryptosporidium (Ruecker et al. 2007; Li et al. 2015)], as well as bacterial markers [Bacteroides HF183/HumM2 and the Enterococcus esp gene (Harwood et al. 2014)]. Bacteroides markers are the most frequently used markers in tracking fecal pollution. Bacteroides is abundant in animal gut, have a short survival time outside its host, and do not replicate after they have been released into the environment (Tsai et al. 2021); therefore, the presence of human Bacteroides markers (i.e., HF183 and HumM2) indicates a more recent human waste exposure. However, current human fecal/sewage-specific markers have advantages and disadvantages. For example, the low concentration of certain viruses and protozoans in some water samples requires the samples to be concentrated for detection (Lee & Kim 2002; Jimenez-Clavero et al. 2005; Wilkes et al. 2009). For some markers, there is a questionable specificity due to their presence in other animal hosts. For example, human bacterial markers such as the Bacteroides-based marker HF183 and HumM2 have been reported to cross-react with canine feces, albeit at lower levels than found in humans (Ahmed et al. 2012; Boehm et al. 2013; Odagiri et al. 2015). Sheludchenko et al. (2011) identified several human-specific Escherichia coli single nucleotide polymorphism (SNP) biomarkers; however, the application of this method requires performing multiple allele-specific real-time PCR, making it labor-intensive. Therefore, a ‘toolbox’ strategy, which integrates several methods, should be used as the most appropriate approach for identifying sources of human fecal/sewage contamination in the water environment (Santo Domingo et al. 2007; Molina et al. 2014).

In a previous study, we have identified naturalized strains of E. coli in wastewater based on the presence of unique SNP biomarker patterns across various intergenic regions of the E. coli genome and for which some of these strains possessed a site-specific insertion of a transposon (IS30) in the uspC-flhDC intergenic region (Zhi et al. 2016a). This uspC-IS30-flhDC marker has only been found in E. coli that has been isolated from wastewater treatment plants (WWTPs). It was not observed in E. coli libraries originating from a wide range of animals (Zhi et al. 2016b, 2019). To determine the unique presence of this marker in wastewater-related E. coli strains, we have searched its sequence against the NCBI database in which thousands of E. coli genomes from various sources (i.e., wastewater, human, animal, food, etc.) had been deposited by researchers across the globe. Interestingly, we found that one E. coli strain carried the uspC-IS30-flhDC marker which was isolated from wastewater in Switzerland, while this marker was not found in E. coli from other sources during the database search (Zhi et al. 2019). To date, we have found these naturalized wastewater strains in all 10 WWTPs in Alberta that we have tested.

We have extensively characterized these strains both phenotypically and genetically (Zhi et al. 2016a, 2017, 2019; Wang et al. 2020), and our data suggest that these strains have evolved to specifically live and survive in a wastewater niche (Zhi et al. 2019). At the whole genome level, these strains are genetically distinct from enteric strains of E. coli (Zhi et al. 2019) and possess a remarkable array of adaptations that allow them to exploit wastewater as a niche. These strains have been shown to be 100 times more resistant to chlorine than some human E. coli strains lacking these markers (Zhi et al. 2017; Wang et al. 2020).

In addition, these naturalized wastewater strains of E. coli are better biofilm producers compared to human fecal strains (Zhi et al. 2017), a mechanism known to be important for survival in harsh environments. Most importantly, these strains all possess the locus of heat resistance (LHR), a transposable genetic element that imparts chlorine, heat, and oxidative resistance to them (Bojer et al. 2010; Mercer et al. 2015; Wang et al. 2020). It was demonstrated that they were 100,000 times more resistant to heat than LHR-negative strains when treated at 60 °C for 5 min. Furthermore, they possess a plethora of genes that are important for survival in a wastewater matrix (Zhi et al. 2019). These genetic/phenotypic properties likely promote the persistence of naturalized wastewater E. coli in the non-host environment. Collectively, our previous studies suggest that these naturalized E. coli strains may represent excellent targets for detecting municipal sewage contamination.

In the present study, we developed a quantitative polymerase chain reaction (qPCR) assay targeting the uspC-IS30-flhDC biomarker of these naturalized wastewater E. coli strains and evaluated the occurrence of these strains across the sewage treatment process. We subsequently compared this E. coli-based marker against the human Bacteroides fecal contamination markers HF183(Bernhard & Field 2000) and HumM2 (Shanks et al. 2009), in addition to other animal fecal contamination markers (cattle, seagulls, and Canada goose), in surface/storm water sources as an exploration of the suitability of this E. coli-based uspC-IS30-flhDC biomarker as a marker of municipal sewage contamination in the environment. The advantage of this assay is that it can be applied to routine water quality monitoring programs that rely on E. coli as a fecal indicator bacteria (FIB) target.

A TaqMan^® qPCR assay was developed to detect the uspC-IS30-flhDC biomarker in naturalized wastewater strains of E. coli. Taqman probe and its corresponding primers were designed using Primer3 (Untergasser et al. 2012). Primers and probes for the uspC-IS30-flhDC marker are shown in Table 1. All reactions were performed in a 20 μl volume containing 5 μl of DNA template, 10 μl of TaqMan^® Fast Advanced Master Mix (Applied Biosystems, Foster City, CA), 900 nM of each primer, and 250 nM of TaqMan probe. Nuclease-free water (Invitrogen, Carlsbad, CA, USA) was used as no template control (NTC). The cycling conditions were 95 °C for 30 s, 40 cycles of 95 °C for 3 s, and 60 °C for 30 s.

To test qPCR sensitivity, a plasmid containing the qPCR target region was constructed. Specifically, genomic DNA from a naturalized wastewater E. coli isolate containing the uspC-IS30-flhDC marker was used as a DNA template, and the PCR was used to amplify the uspC-IS30-flhDC intergenic region (primer set ZIS-F and ZIS-R in Table 1). The PCR product was resolved in a 2% agarose gel in 1× TAE buffer (Promega, Madison, Wisconsin) at 140 V for 45 min and purified from the gel using a QIAquick Gel Extraction Kit (QIAGEN, Inc., Valencia, CA). The amplicon was cloned using a TOPO^® TA Cloning Kit according to manufacturer's instructions (Invitrogen, Inc., Carlsbad, CA). The isolation of recombinant plasmid DNA was performed using the QIAprep Miniprep Kit (QIAGEN, Inc., Valencia, CA). The presence of the correct insert was confirmed by PCR screening and DNA sequencing of the cloned inserts. DNA sequencing was performed by the Applied Genomic Core (TAGC) facilities at the University of Alberta, Edmonton, Canada.

qPCR sensitivity was tested using cloned plasmid DNA constructs. The concentration of plasmid was quantified using a Qubit^® 2.0 Fluorometer (Invitrogen, Carlsbad, CA). The length of the plasmid is 3,908 bp without the insertion of the uspC-IS30-flhDC marker. The constructed plasmid copy number was converted from its concentration based on its molecular weight (Prediger 2017). Ten-fold serial dilutions, in replicates of 12, of plasmid DNA were made and the uspC-IS30-flhDC marker amplified by the qPCR. The 10-fold serial dilutions of plasmid ranged from 100,000 to 0.1 copies/μl. The limit of detection [with 95% confidence intervals (LOD₉₅)] was calculated using the method developed by Wilrich & Wilrich (2009). The specificity of the qPCR was tested against 70 wastewater E. coli strains (41 uspC-IS30-flh positive and 29 uspC-IS30-flh negative), which had been confirmed by sequencing of the intergenic uspC-flhDC region in a previous study (Zhi et al. 2016a).

To determine the specificity of the naturalized strains to sewage and wastewater, we compared their occurrence in wastewater against E. coli-contaminated surface water and groundwater samples across Alberta. One hundred ml of water were processed for E. coli using the Colilert^® presence/absence test according to standard operating procedures in the Alberta Provincial Laboratory for Public Health (ProvLab), Edmonton, Alberta, Canada. ProvLab is the centralized water testing facility in Alberta, which is accredited to the International Standard Organization (ISO) 17025 requirements. In total, 71 E. coli-positive surface water samples and 57 E. coli-positive groundwater samples that were submitted to ProvLab for routine testing from July 2015 to November 2015 were included in this study. Seventy-six wastewater samples (post-grit, secondary-treated, and UV-treated) were collected at two WWTPs located in Calgary, Alberta from February 2015 to May 2015. The wastewater samples were shipped to ProvLab on ice and cultured by Colilert^® for 24 h upon arrival. From all Colilert^® culture-positive water samples incubated for 24 h, 1 ml of culture liquid was removed from the vessel after vigorous mixing and centrifuged at 5,000 × g for 10 min to collect bacterial pellets, and DNA was extracted by using a DNeasy Blood & Tissue kit (QIAGEN, Inc., Valencia, CA). The qPCR conditions for detecting the uspC-IS30-flhDC marker in these samples were the same as those outlined in the previous section.

The existence of PCR inhibitors in these water samples was tested by the amplification of an internal control (IAC) as described by Deer et al. (2010). Specifically, the 198 bp DNA sequence was synthesized and cloned by Integrated DNA Technologies (IDT). The generated internal control plasmid DNA was transformed into E. coli and purified using a QIAprep Miniprep Kit (QIAGEN, Inc., Valencia, CA). Each qPCR of the PCR inhibition test contained 12.5 μl of TaqMan^® Fast Advanced Master Mix (Applied Biosystems, Foster City, CA), 400 nM of each primer, 100 nM of TaqMan probe, 5 μl of DNA template, 5 μl of internal control (100 copies/5 μl), and molecular biology grade water to a total volume of 25 μl. For the IAC reference sample, the 5 μl DNA template was replaced by 5 μl molecular biology grade water. The cycling conditions were 50 °C for 2 min, 95 °C for 30 s, 40 cycles of 95 °C for 3 s, and 60 °C for 30 s. The qPCR threshold was set to 0.2. DNA extracts were considered inhibited if their IAC C_t value was ≥3 cycles higher than the C_t value of the IAC reference sample.

Ninety-three water samples collected from various surface water and storm water systems in southern Alberta, Canada, were used to compare the correlation between the uspC-IS30-flhDC marker and two human markers of fecal sewage pollution (Bacteroides HF183 and HumM2). Additional animal fecal contamination markers [cow (CowM3), seagull (LeeSG), and Canada goose (CGO1)] were also tested. The probe for the CowM3 assay was redesigned using the Primer Express software (Applied Biosystems, Foster City, CA) based on a previous assay by Shanks et al. (2008). The primers and probes used for all markers are provided in Table 1.

Water samples (100 ml) were filtered onto 0.4 μm polycarbonate filters followed by DNA extraction using a PowerWater^® DNA Isolation Kit (Mo Bio Inc., Carlsbad, CA) according to the manufacturer's instructions. The final volume of DNA extraction was 100 μl. All qPCRs were performed in a 20 μl volume containing 5 μl of templates and 10 μl of TaqMan^® Fast Advanced Master Mix (Applied Biosystems, Foster City, CA). The final primer and probe concentrations for each qPCR are shown in Supplementary Material, Table S1. Nuclease-free water (Invitrogen, Carlsbad, CA, USA) was used as NTC. The cycling conditions for qPCR of all markers were 95 °C for 30 s, 40 cycles of 95 °C for 3 s, and 60 °C for 30 s. These water samples were also tested for the existence of PCR inhibitors using the IAC control as described in the previous section.

Statistical tests were performed using R software Version 3.0.0. Cohen's kappa test was used to compare the association between the marker uspC-IS30-flhDC and other human and animal makers. A value of p < 0.05 was considered statistically significant.

A qPCR for the detection of the uspC-IS30-flhDC was developed based on our previous study (Zhi et al. 2016a). The qPCR was designed to target the site-specific location of the IS30 element in the uspC-flhDC intergenic region by targeting the forward primer in the uspC gene, while the reverse primer targeted the IS30 sequence. The qPCR amplicon size was 150 bp.

A typical standard curve, based on plasmid copy numbers ranging from 500,000 to 5, is shown in Supplementary Material, Figure 1. The LOD with 95% confidence (LOD₉₅) for the qPCR was 3.0 × 10⁻⁸ ng of plasmid DNA template, with a 95% confidence interval (CI₉₅) of 1.7 × 10⁻⁸–6.0 × 10⁻⁸ ng. This qPCR was tested against 70 E. coli isolates that were either negative or positive for the uspC-IS30-flhDC marker, as characterized in our previous study (Zhi et al. 2016a). All 41 E. coli strains possessing the uspC-IS30-flhDC marker were positive by qPCR, whereas all 29 E. coli strains lacking this marker were negative by the qPCR assay. In general, our uspC-IS30-flhDC qPCR assay was able to detect the naturalized wastewater E. coli strains at a low LOD with 100% specificity and sensitivity.

The presence of naturalized strains of E. coli possessing the uspC-IS30-flhDC marker was tested in 76 wastewater samples (post-grit removal as well as primary, secondary, and UV-treated), 71 E. coli-contaminated surface water samples, and 57 E. coli-contaminated groundwater samples in order to evaluate the specificity of the uspC-IS30-flhDC qPCR assay to E. coli populations found in wastewater (as opposed to populations found in groundwater and surface water in general). For this comparison, E. coli-cultured (Colilert^®) samples from each of these sources were used for comparison. No PCR inhibitors were observed in these water samples as demonstrated by the PCR inhibition test.

As shown in Table 2, 89% (68/76) of the E. coli-cultured wastewater samples (treated and untreated) were positive for the uspC-IS30-flhDC marker by the qPCR. Among these wastewater samples, 100% of untreated sewage [post-grit removal (31/31)], 100% of primary-treated wastewater (21/21), 92% of secondary-treated wastewater (23/25), and 70% of UV-treated wastewater (14/20) samples were culture positive for the uspC-IS30-flhDC marker (Table 2). By comparison, only 4% of the 71 E. coli culture-positive surface water samples were also positive by the uspC-IS30-flhDC qPCR, and only 5% of the 57 E. coli culture-positive groundwater samples were also found to be positive by the uspC-IS30-flhDC qPCR.

Compared to the high prevalence of the uspC-IS30-flhDC marker in wastewater samples, the much lower prevalence of the marker in E. coli-contaminated surface water and groundwater sources provides additional evidence to our previous findings that the uspC-IS30-flhDC marker-positive E. coli strains might preferentially survive and live in the wastewater environment (Zhi et al. 2016a). Moreover, in the surface and groundwater samples that were positive for the uspC-IS30-flhDC marker, we could not exclude the possibility of the contamination of these water sources with wastewater/sewage, as some surface water sites were known to be downstream of WWTPs and storm drains from urban centers, and it is possible that private groundwater samples may have been impacted by septic systems.

The 57 E. coli culture-positive groundwater drinking well samples were collected through routine water quality monitoring programs in the province of Alberta. Although the sample size may initially seem small, the overall prevalence of E. coli-contaminated groundwater wells in the province of Alberta (in any given year) is only about 2–3% (Invik et al. 2017). Thus, it is estimated that >1,900 groundwater wells were screened by routine public health testing (i.e., culture-based detection of E. coli using in Colilert^®) in order to identify the 57 wells contaminated with E. coli. Using Bacteroides-based assays (HF183 and HumM2) to evaluate human fecal pollution in these same samples would have required all 1,900 groundwater wells to be filtered, DNA extracted and subject to qPCR for each of the targets – a costly endeavor. The distinct advantage of the uspC-IS30-flhDC qPCR assay is that it can be performed on E. coli culture-positive samples only creating a very cost-efficient screening process to identify groundwater or surface water sources contaminated with human sewage/wastewater. For these reasons, the uspC-IS30-flhDC marker appears to be a promising E. coli-based marker for human wastewater contamination.

The qPCR inhibition test of the 93 surface and stormwater samples showed no qPCR inhibition. As a potential marker for municipal sewage contamination, we also evaluated how the uspC-IS30-flhDC E. coli marker compared with the human fecal contamination marker HF183 and the HumM2 in 93 surface water and stormwater samples. In addition, we assessed relationships between the uspC-IS30-flhDC E. coli marker and a variety of animal fecal source-tracking markers in these same samples. The results are shown in Table 3.

The percentage of surface and storm water samples that were positive for HF183 and HumM2 was 61.3% (57/93) and 38.7% (36/93), respectively. By comparison, only 18.3% (17/93) of the samples were positive for uspC-IS30-flhDC. However, among the 17 uspC-IS30-flhDC-positive samples, 16 out of 17 were positive for HF183 and 15 of 17 were also positive for HumM2. All 15 samples that were HumM2/HF183 positive were also positive for uspC-IS30-flhDC. Statistical analysis demonstrated a significant correlation (p < 0.05) between the uspC-IS30-flhDC marker and HF183 and HumM2 markers, suggesting that there are good agreements between these markers in detecting human fecal contamination (Table 3).

Boehm et al. (2015) performed a microbial risk assessment and determined that recreational waters whose HF183 and HumM2 markers’ concentrations were above 4,200 and 2,800 copies/100 ml, respectively, were considered unacceptable by U.S. EPA's criteria for controlling the risk of gastrointestinal illness. When we compared uspC-IS30-flhDC-positive samples with those containing greater than 4,200 copies of human HF183 per 100 ml of water – the upper value of which is considered important for human health risk assessments for swimming in sewage contaminated waters (Boehm et al. 2015) – the discordance between the HF183 and the uspC-IS30-flhDC markers decreased substantially, to only 5.4% (5/93), and with the statistical p-value decreasing considerably (Table 3). Similarly, at a human health risk target of 2,800 copies/100 ml for the human Bacteroides HumM2 target (Boehm et al. 2015), the discordance between the HumM2 and the uspC-IS30-flhDC markers decreased from 24.7% (23/93) to 7.5% (7/93) and was also accompanied by a substantial decrease in the p-value (Table 3). This suggests that the uspC-IS30-flhDC marker may be an extremely valuable tool for screening water of unacceptable health risk in some situations.

These 93 water samples were also tested for three additional animal markers [cow (CowM3), seagull (LeeSG), and Canada goose (CGO1)] (Table 3). Two water samples were found positive for the CowM3 marker. One of the two samples was positive for uspC-IS30-flhDC, but this same sample was also positive for HF183 and HumM2, suggesting that the water sample contained fecal contamination originating from both human/wastewater and cattle. Among the 27 seagull positive samples, five samples were positive for uspC-IS30-flhDC, but these same five samples were also positive for HF183 and four were positive for the HumM2 marker, suggesting that these samples contained both seagull and human/wastewater fecal contamination. In the 11 samples for which the Canada goose marker was observed, only one sample was positive for uspC-IS30-flhDC, and this sample was also positive for both human HF183 and HumM2. Overall, statistical analysis (Table 3) showed that no agreement was observed between uspC-IS30-flhDC and the cow marker CowM3 (p = 0.24), seagull marker LeeSG (p = 1), and the Canada goose marker COG1 (p = 0.40) in detecting human fecal contaminations.

At present, several human fecal/sewage-specific indicators have been characterized, including human-specific viruses [e.g., adenovirus, enterovirus, and polyomavirus (Jiang et al. 2001; van der Sanden et al. 2013; Rusinol et al. 2016)], protozoan parasites [e.g., Cryptosporidium (Ruecker et al. 2007; Li et al. 2015)], as well as bacterial markers [Bacteroides HF183/HumM2 and the Enterococcus esp gene (Harwood et al. 2014)]. However, each of these markers has advantages and disadvantages. For example, the low concentration of certain viruses and protozoans in some water samples requires the samples to be concentrated for detection (Lee & Kim 2002; Jimenez-Clavero et al. 2005; Wilkes et al. 2009). For some markers, there is a questionable specificity due to their presence in other animal hosts. For example, human bacterial markers such as the Bacteroides-based marker HF183 and HumM2 have been reported to cross-react with canine feces, albeit at lower levels than found in humans (Ahmed et al. 2012; Boehm et al. 2013; Odagiri et al. 2015). Therefore, a ‘toolbox’ strategy, which integrates several methods, has been proposed as the most appropriate approach for identifying sources of human fecal/sewage contamination in the water environment (Santo Domingo et al. 2007; Molina et al. 2014). Our results have demonstrated that the qPCR assay of the uspC-IS30-flhDC marker might be a potential tool in detecting human fecal/wastewater contamination. At present, culture-based assays for E. coli remain the gold standard for microbial water quality testing. Therefore, our qPCR method for detecting naturalized wastewater E. coli from cultured samples is easily adaptable to current regulatory monitoring programs. This provides the added advantage of determining whether human sewage is contaminating the water source.

Although we have demonstrated that uspC-IS30-flhDC-positive E. coli strains seem specifically adapted to and survive in wastewater through the extensive NCBI online database search and testing again a local library of 780 E. coli strains from 15 animal hosts (Zhi et al. 2016a, 2019), future study is still needed to extensively evaluate the presence of this uspC-IS30-flhDC marker in non-human sewage contaminated samples. For instance, non-human fecal material (i.e., cattle, chicken, and dog) could be used to determine the true negative rate of this maker before integrating this marker in routine analysis of human wastewater contamination.

In summary, we have developed a qPCR assay targeting the uspC-IS30-flhDC marker of naturalized wastewater E. coli. The uspC-IS30-flhDC marker showed strong correlation to the human fecal markers Bacteroides HF183 and HumM2, but not animal fecal markers in surface and stormwater samples. The uspC-IS30-flhDC marker appears to be a promising E. coli-based marker for human wastewater contamination and adaptable to routine water quality monitoring programs using E. coli as FIB.

We would like to thank the Environmental Microbiology staff at the Provincial Laboratory for Public Health (Calgary), in particular Lorraine Ingham, for overseeing the water sample analysis for routine E. coli culture and the pre-analytical processing of these samples for alternative markers. We would also like to thank Dr Norma Ruecker, Dr Theingi Maw, Jennifer Berwanger, and the rest of the City of Calgary Water Quality Services for organizing the sampling of WWTPs at the City of Calgary and performing culture-based quantitation of E. coli from wastewater samples.

This work was supported by Alberta Innovates, the City of Calgary, the Natural Sciences and Engineering Research Council (NSERC), the Canadian Foundation for Innovation (CFI), the National Natural Sciences Foundation of China (No. 82073514), and the Natural Science Foundation of Ningbo (No. 202003N4114).

The authors declare that they have no competing interests.

All relevant data are included in the paper or its Supplementary Information.