Testing of General and Human-Associated Fecal Contamination in Waters.

qPCR has become increasingly popular for microbial water quality testing because it is faster, more specific, and more flexible than culture-based methods. However, qPCR method limitations such as quantification bias introduced by reliance on standards and susceptibility to PCR inhibitors are major obstacles for implementation in water testing. This is because water testing requires accurate quantification of rare targets and because environmental waters often contain PCR inhibitors. Digital PCR offers the opportunity to maintain qPCR's advantages over culture-based methods while ameliorating two of qPCR's major limitations: the necessity to run standard curves and high susceptibility to inhibition. Here we describe a complete method for simultaneous testing for a general microbial water quality indicator (Enterococcus spp.) and a human-associated fecal marker in environmental waters. The complete method includes water sampling and filtration to capture bacteria, DNA extraction from bacteria captured on the filter, and droplet digital PCR to quantify the genetic markers from bacteria indicative of general and human-associated fecal contamination.

Regional Assessment of Human Fecal Contamination in Southern California Coastal Drainages.

Host-associated genetic markers that allow for fecal source identification have been used extensively as a diagnostic tool to determine fecal sources within watersheds, but have not been used in routine monitoring to prioritize remediation actions among watersheds. Here, we present a regional assessment of human marker prevalence among drainages that discharge to the U.S. southern California coast. Approximately 50 samples were analyzed for the HF183 human marker from each of 22 southern California coastal drainages under summer dry weather conditions, and another 50 samples were targeted from each of 23 drainages during wet weather. The HF183 marker was ubiquitous, detected in all but two sites in dry weather and at all sites during wet weather. However, there was considerable difference in the extent of human fecal contamination among sites. Similar site ranking was produced regardless of whether the assessment was based on frequency of HF183 detection or site average HF183 concentration. However, site ranking differed greatly between dry and wet weather. Site ranking also differed greatly when based on enterococci, which do not distinguish between pollution sources, vs. HF183, which distinguishes higher risk human fecal sources from other sources, indicating the additional value of the human-associated marker as a routine monitoring tool.

A Duplex Digital PCR Assay for Simultaneous Quantification of the Enterococcus spp. and the Human Fecal-associated HF183 Marker in Waters.

This manuscript describes a duplex digital PCR assay (EntHF183 dPCR) for simultaneous quantification of Enterococcus spp. and the human fecal-associated HF183 marker. The EntHF183 duplex dPCR (referred as EntHF183 dPCR hereon) assay uses the same primer and probe sequences as its published individual quantitative PCR (qPCR) counterparts. Likewise, the same water filtration and DNA extraction procedures as performed prior to qPCR are followed prior to running dPCR. However, the duplex dPCR assay has several advantages over the qPCR assays. Most important, the dPCR assay eliminates the need for running a standard curve and hence, the associated bias and variability, by direct quantification of its targets. In addition, while duplexing (i.e., simultaneous quantification) Enterococcus and HF183 in qPCR often leads to severe underestimation of the less abundant target in a sample, dPCR provides consistent quantification of both targets, whether quantified individually or simultaneously in the same reaction. The dPCR assay is also able to tolerate PCR inhibitor concentrations that are one to two orders of magnitude higher than those tolerated by qPCR. These advantages make the EntHF183 dPCR assay particularly attractive because it simultaneously provides accurate and repeatable information on both general and human-associated fecal contamination in environmental waters without the need to run two separate qPCR assays. Despite its advantages over qPCR, the upper quantification limit of the dPCR assay with currently available instrumentation is approximately four orders of magnitude lower than that achievable by qPCR. Consequently, dilution is needed for measurement of high concentrations of target organisms such as those typically observed following sewage spills.

Multi-laboratory survey of qPCR enterococci analysis method performance in U.S. coastal and inland surface waters.

Quantitative polymerase chain reaction (qPCR) has become a frequently used technique for quantifying enterococci in recreational surface waters, but there are several methodological options. Here we evaluated how three method permutations, type of mastermix, sample extract dilution and use of controls in results calculation, affect method reliability among multiple laboratories with respect to sample interference. Multiple samples from each of 22 sites representing an array of habitat types were analyzed using EPA Method 1611 and 1609 reagents with full strength and five-fold diluted extracts. The presence of interference was assessed three ways: using sample processing and PCR amplifications controls; consistency of results across extract dilutions; and relative recovery of target genes from spiked enterococci in water sample compared to control matrices with acceptable recovery defined as 50 to 200%. Method 1609, which is based on an environmental mastermix, was found to be superior to Method 1611, which is based on a universal mastermix. Method 1611 had over a 40% control assay failure rate with undiluted extracts and a 6% failure rate with diluted extracts. Method 1609 failed in only 11% and 3% of undiluted and diluted extracts analyses. Use of sample processing control assay results in the delta-delta Ct method for calculating relative target gene recoveries increased the number of acceptable recovery results. Delta-delta tended to bias recoveries from apparent partially inhibitory samples on the high side which could help in avoiding potential underestimates of enterococci--an important consideration in a public health context. Control assay and delta-delta recovery results were largely consistent across the range of habitats sampled, and among laboratories. The methodological option that best balanced acceptable estimated target gene recoveries with method sensitivity and avoidance of underestimated enterococci densities was Method 1609 without extract dilution and using the delta-delta calculation method. The applicability of this method can be extended by the analysis of diluted extracts to sites where interference is indicated but, particularly in these instances, should be confirmed by augmenting the control assays with analyses for target gene recoveries from spiked target organisms.

Droplet digital PCR for simultaneous quantification of general and human-associated fecal indicators for water quality assessment.

Despite wide application to beach water monitoring and microbial source identification, results produced by quantitative PCR (qPCR) methods are subject to bias introduced by reliance on quantitative standards. Digital PCR technology provides direct, standards-free quantification and may potentially alleviate or greatly reduce other qPCR limitations such as difficulty in multiplexing and susceptibility to PCR inhibition. This study examined the efficacy of employing a duplex droplet digital PCR (ddPCR) assay that simultaneously quantifies Enterococcus spp. and the human fecal-associated HF183 marker for water quality assessment. Duplex ddPCR performance was evaluated side-by-side with qPCR and simplex ddPCR using reference material and 131 fecal and water samples. Results for fecal and water samples were highly correlated between ddPCR and simplex qPCR (coefficients > 0.93, p < 0.001). Duplexing Enterococcus and HF183 in qPCR led to competition and resulted in non-detection or underestimation of the target with low concentration relative to the other, while results produced by simplex and duplex ddPCR were consistent and often indistinguishable from one another. ddPCR showed greater tolerance for inhibition, with no discernable effect on quantification at inhibitor concentrations one to two orders of magnitude higher than that tolerated by qPCR. Overall, ddPCR also exhibited improved precision, higher run-to-run repeatability, similar diagnostic sensitivity and specificity on the HF183 marker, but a lower upper limit of quantification than qPCR. Digital PCR has the potential to become a reliable and economical alternative to qPCR for recreational water monitoring and fecal source identification. Findings from this study may also be of interest to other aspects of water research such as detection of pathogens and antibiotic resistance genes.

Detection limits and cost comparisons of human- and gull-associated conventional and quantitative PCR assays in artificial and environmental waters.

Some molecular methods for tracking fecal pollution in environmental waters have both PCR and quantitative PCR (qPCR) assays available for use. To assist managers in deciding whether to implement newer qPCR techniques in routine monitoring programs, we compared detection limits (LODs) and costs of PCR and qPCR assays with identical targets that are relevant to beach water quality assessment. For human-associated assays targeting Bacteroidales HF183 genetic marker, qPCR LODs were 70 times lower and there was no effect of target matrix (artificial freshwater, environmental creek water, and environmental marine water) on PCR or qPCR LODs. The PCR startup and annual costs were the lowest, while the per reaction cost was 62% lower than the Taqman based qPCR and 180% higher than the SYBR based qPCR. For gull-associated assays, there was no significant difference between PCR and qPCR LODs, target matrix did not effect PCR or qPCR LODs, and PCR startup, annual, and per reaction costs were lower. Upgrading to qPCR involves greater startup and annual costs, but this increase may be justified in the case of the human-associated assays with lower detection limits and reduced cost per sample.

Multi-laboratory evaluations of the performance of Catellicoccus marimammalium PCR assays developed to target gull fecal sources.

Here we report results from a multi-laboratory (n = 11) evaluation of four different PCR methods targeting the 16S rRNA gene of Catellicoccus marimammalium originally developed to detect gull fecal contamination in coastal environments. The methods included a conventional end-point PCR method, a SYBR(®) Green qPCR method, and two TaqMan(®) qPCR methods. Different techniques for data normalization and analysis were tested. Data analysis methods had a pronounced impact on assay sensitivity and specificity calculations. Across-laboratory standardization of metrics including the lower limit of quantification (LLOQ), target detected but not quantifiable (DNQ), and target not detected (ND) significantly improved results compared to results submitted by individual laboratories prior to definition standardization. The unit of measure used for data normalization also had a pronounced effect on measured assay performance. Data normalization to DNA mass improved quantitative method performance as compared to enterococcus normalization. The MST methods tested here were originally designed for gulls but were found in this study to also detect feces from other birds, particularly feces composited from pigeons. Sequencing efforts showed that some pigeon feces from California contained sequences similar to C. marimammalium found in gull feces. These data suggest that the prevalence, geographic scope, and ecology of C. marimammalium in host birds other than gulls require further investigation. This study represents an important first step in the multi-laboratory assessment of these methods and highlights the need to broaden and standardize additional evaluations, including environmentally relevant target concentrations in ambient waters from diverse geographic regions.

Evaluation of the repeatability and reproducibility of a suite of qPCR-based microbial source tracking methods.

Many PCR-based methods for microbial source tracking (MST) have been developed and validated within individual research laboratories. Inter-laboratory validation of these methods, however, has been minimal, and the effects of protocol standardization regimes have not been thoroughly evaluated. Knowledge of factors influencing PCR in different laboratories is vital to future technology transfer for use of MST methods as a tool for water quality management. In this study, a blinded set of 64 filters (containing 32 duplicate samples generated from 12 composite fecal sources) were analyzed by three to five core laboratories with a suite of PCR-based methods utilizing standardized reagents and protocols. Repeatability (intra-laboratory variability) and reproducibility (inter-laboratory variability) of observed results were assessed. When standardized methodologies were used, intra- and inter-laboratory %CVs were generally low (median %CV 0.1-3.3% and 1.9-7.1%, respectively) and comparable to those observed in similar inter-laboratory validation studies performed on other methods of quantifying fecal indicator bacteria (FIB) in environmental samples. ANOVA of %CV values found three human-associated methods (BsteriF1, BacHum, and HF183Taqman) to be similarly reproducible (p > 0.05) and significantly more reproducible (p < 0.05) than HumM2. This was attributed to the increased variability associated with low target concentrations detected by HumM2 (approximately 1-2 log10copies/filter lower) compared to other human-associated methods. Cow-associated methods (BacCow and CowM2) were similarly reproducible (p > 0.05). When using standardized protocols, variance component analysis indicated sample type (fecal source and concentration) to be the major contributor to total variability with that from replicate filters and inter-laboratory analysis to be within the same order of magnitude but larger than inherent intra-laboratory variability. However, when reagents and protocols were not standardized, inter-laboratory %CV generally increased with a corresponding decline in reproducibility. Overall, these findings verify the repeatability and reproducibility of these MST methods and highlight the need for standardization of protocols and consumables prior to implementation of larger scale MST studies involving multiple laboratories.

Comparison of PCR and quantitative real-time PCR methods for the characterization of ruminant and cattle fecal pollution sources.

The State of California has mandated the preparation of a guidance document on the application of fecal source identification methods for recreational water quality management. California contains the fifth highest population of cattle in the United States, making the inclusion of cow-associated methods a logical choice. Because the performance of these methods has been shown to change based on geography and/or local animal feeding practices, laboratory comparisons are needed to determine which assays are best suited for implementation. We describe the performance characterization of two end-point PCR assays (CF128 and CF193) and five real-time quantitative PCR (qPCR) assays (Rum2Bac, BacR, BacCow, CowM2, and CowM3) reported to be associated with either ruminant or cattle feces. Each assay was tested against a blinded set of 38 reference challenge filters (19 duplicate samples) containing fecal pollution from 12 different sources suspected to impact water quality. The abundance of each host-associated genetic marker was measured for qPCR-based assays in both target and non-target animals and compared to quantities of total DNA mass, wet mass of fecal material, as well as Bacteroidales, and enterococci determined by 16S rRNA qPCR and culture-based approaches (enterococci only). Ruminant- and cow-associated genetic markers were detected in all filters containing a cattle fecal source. However, some assays cross-reacted with non-target pollution sources. A large amount of variability was evident across laboratories when protocols were not fixed suggesting that protocol standardization will be necessary for widespread implementation. Finally, performance metrics indicate that the cattle-associated CowM2 qPCR method combined with either the BacR or Rum2Bac ruminant-associated methods are most suitable for implementation.

Bacteroidales terminal restriction fragment length polymorphism (TRFLP) for fecal source differentiation in comparison to and in combination with universal bacteria TRFLP.

Terminal restriction fragment length polymorphism (TRFLP) is an attractive community analysis method for microbial source tracking (MST) because it is accessible, relatively inexpensive, and can discern multiple fecal sources simultaneously. A new Bacteroidales TRFLP (Bac-TRFLP) method was developed and its source identification performance was evaluated by itself, in comparison to, and in combination with an existing universal bacterial TRFLP method in two laboratories. Sixty-four blind samples from 12 fecal sources (sewage, septage, human, dog, horse, cow, deer, pig, chicken, goose, pigeon, and gull) were used for evaluation. Bac- and Univ-TRFLP exhibited similarly high overall correct identification (>88% and >89%, respectively), excellent specificity regardless of fecal sources, variable sensitivity depending on the source, and stable performance across two laboratories. Compared to Univ-TRFLP, Bac-TRFLP had better sensitivity and specificity with horse, cow, and pig fecal sources but was not suited for certain avian sources such as goose, gull, and pigeon. Combining the general and more targeted TRFLP methods (Univ&Bac-TRFLP) achieved higher overall correct identification (>92%), higher sensitivity and specificity metrics, and higher reproducibility between laboratories. Our results suggest that the Bac-TRFLP and Univ&Bac-TRFLP methods are promising additions to the MST toolbox and warrant further evaluation and utilization in field MST applications.