Correlative Assessment of Fecal Indicators using Human

Article pubs.acs.org/est

Correlative Assessment of Fecal Indicators using Human Mitochondrial DNA as a Direct Marker Vikram Kapoor,† Christopher Smith,† Jorge W. Santo Domingo,‡ Ting Lu,§ and David Wendell*,† †

School of Energy, Environmental, Biological & Medical Engineering, University of Cincinnati, Cincinnati, Ohio 45221, United States Office of Research and Development, U.S. Environmental Protection Agency, Cincinnati, Ohio 45221, United States § Metropolitan Sewer District of Greater Cincinnati, Cincinnati, Ohio 45204, United States ‡

S Supporting Information *

ABSTRACT: Identifying the source of surface water fecal contamination is paramount to mitigating pollution and risk to human health. Fecal bacteria such as E. coli have been staple indicator organisms for over a century, however there remains uncertainty with E. coli-based metrics since these bacteria are abundant in the environment. The relationships between the presence of direct indicator of human waste (human mitochondrial DNA), human-specific Bacteroidales, and E. coli were studied for water samples taken from an urban creek system (Duck Creek Watershed, Cincinnati, OH) impacted by combined sewer overflows. Logistic regression analysis shows that human-specific Bacteroidales correlates much more closely to human mitochondrial DNA (R = 0.62) relative to E. coli (R = 0.33). We also examine the speciation of Bacteroidales within the Duck Creek Watershed using next-generation sequencing technology (Ion Torrent) and show the most numerous populations to be associated with sewage. Here we demonstrate that human-specific Bacteroidales closely follow the dynamics of human mitochondrial DNA concentration changes, indicating that these obligate anaerobes are more accurate than E. coli for fecal source tracking, lending further support to risk overestimation using coliforms, especially fecal coliforms and E. coli.

■

INTRODUCTION Water quality deterioration due to fecal contamination caused by human and animal sources is a serious concern for many countries in the world.1−3 High levels of fecal bacteria are the primary cause of river and stream impairment in the United States according to the National Water Quality Inventory,4 with numerous contamination sources often uncharacterized.1 These include municipal waste from household sewage treatment systems, combined sewer overflows (CSO), sanitary sewer overflows and other factors (agriculture/urban runoff, water temperature and available organic material).5,6 Hence, accurate and reliable fecal source identification methods are essential for mitigating bacterial contributions to waterways and maintaining water quality. Methods to determine the sources of fecal contamination in environmental waters have evolved and grown over many years.7,8 There is presently no single biological or chemical fecal identifier used in all water systems. Various approaches have been used to identify fecal sources in water samples ranging from microbiological, phenotypic and genotypic methods to chemical analysis. Chemical makers used to detect humanassociated waste include fecal sterols, caffeine, fragrances, detergents, pharmaceuticals and personal care products;9−11 however, issues with persistence and detection of these chemicals limits their use as reliable source identification tools.10 Escherichia coli and Enterococci remain the most widely used indicators for fecal pollution because they exist in the intestinal © XXXX American Chemical Society

tract of humans and other animals in large numbers. However, these conventional fecal indicators are unreliable for MST applications due to widely varying survival rates in environment,12,13 growth in ambient or nonenteric habitats,14,15 ability to replicate after discharge in water,16,17 low levels of correlation with the presence of pathogens,18,19 and inability to distinguish between fecal bacteria associated with recent contamination events and those adapted to secondary habitats.20,21 A variety of recent investigations seek to overcome the above limitations by targeting the Bacteroidales 16S rRNA using molecular techniques such as polymerase chain reaction (PCR) or a quantitative PCR-based one.22−25 Members of the Bacteroidales group are abundant in animal feces26,27 and are considered useful fecal indicators since most are obligate anaerobes presumed to survive for only short periods of time after release from their hosts.28,29 As a result, Bacteroidales do not have the ambiguous survival and sourcing issues associated with E. coli and Enterococcus sp., and can be linked with good spatiotemporal resolution to a sampling site.22,24 PCR amplification of host-specific Bacteroidales 16S rRNA gene sequences has been used by Kildare et al.22 for quantitative detection of universal, human-, ruminant-, and dog-specific Received: May 7, 2013 Revised: July 28, 2013 Accepted: August 6, 2013

A

dx.doi.org/10.1021/es4020458 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 1. Locations of sampling sites (highlighted in pink).

fecal Bacteroidales. Recently, additional PCR-based assays have been developed specifically for human-associated Bacteroidales marker,30,31 swine-associated Bacteroidales marker,32 chickenand duck-associated Bacteroides marker,33 and even phages infecting Bacteroides.34 To date, few studies have determined the efficacy of using Bacteroides 16S rRNA genetic markers as fecal indicators for the occurrence of bacterial pathogens.18,35 Arguably, a more direct approach would be to relate the presence of Bacteroides markers to the host species-specific DNA sequences in the form of mitochondrial DNA (mtDNA). Mitochondrial genome contains species-specific sequences which can be used in a similar manner to microbial markers for fecal source tracking with source identification advantages which exceed bacterial indicators.36 This approach is important because associated bacteria may have come from alternative hosts or may be surviving in the stream, making the temporal component of contamination ambiguous. Thus, PCR of mtDNA can be used

to identify animal waste directly through its own discharged eukaryotic cells.36 Feces from different sources, composing human fecal waste and waste from domesticated animals, contain large amounts of exfoliated epithelial cells, and mtDNA has hundreds to thousands of copies per cell.37 Therefore, mtDNA provides a robust PCR target similar to 16S rRNA genes, with detectable DNA persisting even after cell death. However, mtDNA is still an indirect indicator of the presence of bacterial pathogens and risk estimation of human health. The application of mitochondrial DNA for fecal source tracking has been demonstrated in the past.36,38,39 Thus, PCR of 16S rRNA genes of the obligate anaerobe Bacteroides sp. along with evidence of mitochondrial DNA as a direct marker of human fecal waste has the potential to be a more reliable way of fecal source tracking for environmental waters, since neither of these replicate in oxygenated surface waters, although alternate habitats such as sediments and decaying vegetation with B



Article

Table 1. PCR and qPCR Primer Sets Used in This Study

■

MATERIALS AND METHODS Sampling Sites. A total of 10 sampling points were selected within the Duck Creek watershed, which were chosen based on proximity to CSO’s, local municipal sampling sites and potential impact from human fecal pollution from CSOs and watershed runoff (Figure 1; also see Supporting Information Table S1). Sites 1, 2, 3, 7, and 10 are located on Duck Creek at river mile 2.0, 2.4, 3.38, 4.5, and 5.0, and sites 4, 5, and 6 are located at three stations on Little Duck Creek; river mile 0.49, 1.7, and 2.2 (Supporting Information Table S2). Sites 8 and 9 are located on Deerfield Creek, near CSO 556, since this overflow had the highest number of annual overflow events and largest ever volumetric contribution to the CSO total overflow. Sites 7 and 10 are close to CSO 68 which had the second highest contribution and lies at the confluence of Upper Duck Creek and Deerfield Creek. A detailed description of the watershed and sampling sites is given in Supporting Information. Sample Collection and DNA Extraction. Water samples (n = 180) were collected monthly over a period of 18 months (July 2011 to December 2012) from 10 different sites within the Duck Creek Watershed (Figure 1). All samples were collected using sterilized 1 L bottles (Nalgene, Rochester, NY) and transported on ice to the laboratory at the University of Cincinnati (Clifton, OH) within 2 h of collection. Water samples (500 mL) were filtered through 0.45-μm-pore-size, 47mm-diameter mixed cellulose ester filters (Millipore, Billerica, MA) and stored at −80 °C until DNA extraction. DNA was extracted from filter samples using the PowerWater DNA Isolation Kit (MO BIO Laboratories, Carlsbad, CA) according

anaerobic conditions do exist under ambient surroundings where these bacteria are likely to grow. Investigations examining the simultaneous occurrence and prevalence of host-specific mitochondrial DNA in environmental waters in relation to fecal indicators are limited. Moreover, there has been no data on the relationships between the occurrence of mitochondrial DNA and alternative (Bacteroidales 16S rRNA genetic markers) and conventional (E. coli) fecal indicators. The Duck Creek watershed (Cincinnati, OH) was used for sampling due to continued fecal contributions to the Little Miami River recreational waterway.40 The correlation between the presence of human mtDNA, human-specific Bacteroidales and E. coli at sampling sites was studied to design an integrated fecal source tracking strategy within segments of the aforementioned watershed. We have applied end-point PCR and quantitative PCR assays with Bacteroidales host-specific primers in conjunction with mitochondrial DNA-based assays to identify and quantify host-specific fecal contamination. Furthermore, we investigated the genetic diversity of Bacteroidales 16S rRNA gene sequences derived from some of the water samples to substantiate the evidence of human wastes in the watershed, and to examine their relative abundance among sampling sites. This information is also necessary to determine which host and environmental sequences were contributing to fecal pollution in environmental waters and to identify populations relevant to human fecal contamination. C



Article

Standard curves were generated using 10-fold serial dilutions (106−101) of genomic DNA copies of Bacteroides f ragilis as well as DNA copies generated from PCR amplifications of humanspecific Bacteroidales 16S rRNA genetic marker, human-specific mitochondrial DNA, and E. coli uidA gene respectively. Standards were created by diluting purified PCR amplicons to known copy numbers. The target copy numbers (T) were estimated by the equation T = [D/(AL × 660)] × 6.022 × 1023, where D (g/μL) is amplicon DNA concentration and AL (in base pairs) is amplicon length. The standard curves for each qPCR assay were then generated by plotting the threshold cycle (Ct) values against the numbers of target copies corresponding to the serially diluted standards. Amplicons were sequenced and NCBI BLAST searches were performed to verify sequence identities. PCR amplification efficiencies were calculated by the instrument manufacturer’s instructions (Applied Biosystems). Internal amplification controls (IAC) were employed to check for PCR inhibitors: known copies of each amplicon were added to a sample aliquot and compared to the relevant standard curve or another IAC sample with only water and master mix added. Data analysis of the qPCR standard curves was performed using Sequence Detection Software version 1.2.3 (Applied Biosystems, Green Island, NY). Bacteroidales 16S rRNA Gene Amplicon Sequencing. The Ion Torrent Personal Genome Machine (PGM) system (Life Technologies, San Francisco, CA) was used to examine the sequence diversity of Bacteroidales populations. Sample multiplexing using unique molecular barcodes (10 nucleotides in length) was employed for combined, cost-effective semiconductor based high throughput sequencing (Supporting Information Table S3). Bacteroidales 16S rRNA gene amplification was carried out using the Bacteroidales-based 16 S rDNA targeting primers BacUni-520f (5′-CGTTATCCGGATTTATTGGGTTTA-3′) and BacUni-690r (5′-CAATCGGAGTTCTTCGTGATATCTA-3′) linked to the barcodes. Reactions were performed in 50 μL volumes using the Taq master mix (New England Biolabs, Ipswich, MA) with 200 nM each of the forward and reverse primer and 5 μL of template DNA. Cycling conditions involved an initial 5 min denaturing step at 95 °C, followed by 30 cycles of 30 s at 95 °C, 30 s at 60 °C, and 30 s at 68 °C and a final elongation step of 5 min at 68 °C. All PCR products were purified using MinElute PCR Purification kit (Qiagen, Valencia, CA) and quantified using a NanoDrop 1000 Spectrophotometer (Thermo Scientific, Wilmington, DE). The amplification products from the different samples were pooled in an equimolar ratio to conduct multiplexed sequencing. Sequencing of the pooled library was performed on the PGM system using a 314 chip or 316 chip with the Ion Sequencing 200 kit according to the manufacturer’s protocol. The quality metrics used were derived from the automated analysis carried out by the Torrent Suite Software version 3.4.2 (Life Technologies). Sequences were compiled and cleaned using CLC Genomics Workbench Version 6.0.2 (CLC Bio, Cambridge, MA). All sequences having an average quality under 17, having unidentified bases (Ns), not exactly matching the barcode sequence, or being shorter than 75 bp were discarded. Sequences were then submitted to BLAST homology search algorithms to assess similarity to sequences in the 16S rRNA sequences (Bacteria) database (NCBI). Statistical Analyses. The correlation between the concentrations of human mtDNA, E. coli and human-specific Bacteroidales 16S rRNA gene marker were analyzed using the

to the manufacturer’s protocol. Extraction controls with purified water were used during filtration to monitor for potential extraneous DNA contamination. The concentration and purity of DNA was determined using NanoDrop 1000 Spectrophotometer (Thermo Scientific, Wilmington, DE). DNA extracts were stored at −20 °C for subsequent analyses. Analyses of Bacteroidales 16S rRNA Genetic Markers. Bacteroidales 16S rRNA gene-based PCR assays were used to detect the presence of Bacteroidales populations as well as human-specific, bovine-specific and canine-specific markers as described by Kildare et al.22 For all PCR assays, water DNA extracts (5 μL) were used as templates in a final reaction volume of 50 μL using the Taq master mix (New England Biolabs, Ipswich, MA) with 200 nM each of the forward and reverse primer in a GeneAmp PCR System 9700 thermal cycler (Applied Biosystems, Green Island, NY) under the following cycling conditions: initial denaturation of 30 s at 95 °C, followed by 35 cycles of 15 s at 95 °C, 20 s at 60 °C and 30 s at 68 °C, and final extension step of 5 min at 68 °C. The PCR products were visualized using E-Gel 2% with SYBR Safe (Invitrogen, Green Island, NY) to detect the presence/absence of all Bacteroidales 16S rRNA gene assays. Controls containing no template DNA were used to check for cross contamination. Additionally, PCR inhibition was tested in water DNA extracts by amplifying with general bacterial 16S rRNA gene-targeted primer sets and by using 10-fold dilutions of each DNA extract. Analyses of Mitochondrial DNA Markers. PCR assays targeting the mitochondrial gene NADH dehydrogenase subunit 5 (ND5) were used to detect human-, bovine- and swine-specific mitochondrial DNA using host-specific primer sets previously designed by Caldwell et al.36 For all PCR assays, water DNA extracts (5 μL) were used as templates in a final reaction volume of 50 μL using the Taq master mix (New England Biolabs, Ipswich, MA) with 200 nM each of the forward and reverse primer in a GeneAmp PCR System 9700 thermal cycler (Applied Biosystems, Green Island, NY) under the following cycling conditions: initial denaturation of 30 s at 95 °C, followed by 35 cycles of 15 s at 95 °C, 20 s at 60 °C and 30 s at 68 °C, and final extension step of 5 min at 68 °C. The PCR products were visualized using E-Gel 2% with SYBR Safe (Invitrogen, Green Island, NY) to detect the presence/absence of host-specific mitochondrial DNA with controls lacking template used to check for cross contamination. qPCR Analyses. SYBR Green-based qPCR assays were performed to determine the relative abundance of humanspecific Bacteroidales 16S rRNA gene marker, human-specific mitochondrial DNA, and E. coli uidA gene (Table 1). The samples were placed in MicroAmp Optical 96-well plates and amplified in a 7500 Real-Time PCR System (Applied Biosystems, Green Island, NY). Each PCR mixture (50 μL) was composed of 25 μL of 1× OneTaq master mix (New England Biolabs, Ipswich, MA), 200 nM of each forward and reverse primers, 0.5X SYBR Green 1 dye, and 5 μL of template DNA. Standard amplification conditions were used: 2 min at 50 °C and 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. No-template controls were used to check for cross contamination, and 10-fold dilutions of each DNA extract were used to test for PCR inhibition. Dissociation curves were examined to determine the presence of potential primer dimers and other nonspecific reaction products. PCR products were also visualized using E-Gel 2% with SYBR Safe (Invitrogen, Green Island, NY) to confirm the size of amplification products. D



Article

Table 2. Distribution of Molecular Markers Used in This Studya % of tested water samples positive for marker (no. of samples tested) site 1 2 3 4 5 6 7 8 9 10 a

BacUni 100 100 100 100 100 100 100 100 100 100

(18) (18) (18) (18) (18) (18) (18) (18) (18) (18)

BacHum 100 100 100 100 100 100 100 100 100 100

(18) (18) (18) (18) (18) (18) (18) (18) (18) (18)

BacCow

BacCan

ND ND ND ND ND ND ND ND ND ND

11 (18) ND 16 (18) ND ND ND 83 (18) 33 (18) 22 (18) 89 (18)

human mt 83 89 89 56 56 78 72 78 89 89

(18) (18) (18) (18) (18) (18) (18) (18) (18) (18)

bovine mt

swine mt



uidA 100 100 100 100 100 100 100 100 100 100

(18) (18) (18) (18) (18) (18) (18) (18) (18) (18)

ND = not detected.

Figure 2. Correlations between human mtDNA and fecal indicators. (A) Human-specific Bacteroidales 16S rRNA genetic markers vs human mtDNA and (B) E. coli uidA gene marker vs human mtDNA (n = 126).

logistic regression tests at α-values of 0.001 and 0.0001. The null hypothesis (H0) was that the regression line slope is equal to zero; in other words, as the concentration of human mtDNA increases, the concentration of the associated fecal indicator gets neither higher nor lower. H0 was rejected when the test result (i.e., p-value) was less than the target α-value (0.001 or 0.0001). All samples subjected to qPCR with a Ct value above background were included for analysis (n = 126). Correlation coefficients (r2) were also determined for the data sets. Linear

regression analysis was also used to verify if differences in concentration between Bacteroidales marker and E. coli bacteria were significant. All analyses were performed using Microsoft Excel (2011) and correlation strength was interpreted according to an accepted scale for biological statistics.41

■

RESULTS AND DISCUSSION Performance of qPCR Assays. Standard curves were generated using serial dilutions of known copy numbers to E



Article

uidA gene marker) in this study (Figure 2). There was a strong correlation between the human-specific Bacteroidales 16S rRNA genetic markers and human mtDNA in the samples (correlation coefficient, r2 = 0.38, p-value < 0.0001) suggesting a similar persistence for mtDNA and human-specific Bacteroidales in environmental waters. The correlation between E. coli uidA gene copies and human mtDNA copies was weaker (r2 = 0.11, p-value < 0.001). The strong correlation of human mtDNA and human-specific Bacteroidales, confirms the value of Bacteroidales as human fecal indicators. This study has for the first time demonstrated a cocorrelation between the presence of this human-specific MST marker and the presence of the human mitochondrial genome in an urban watershed. Previously, it has been demonstrated that phages infecting Bacteroides GB-124 correlate positively with the presence of human enteric viruses, such as human adenoviruses and noroviruses,34 which in turn support the observation that Bacteroides spp. are an important indicator of human fecal contamination. E. coli counts were higher than those of human-specific Bacteroidales 16S rRNA marker determined by qPCR for almost all surface water samples. This is expected based on the difference in survival and persistence of E. coli and anaerobic Bacteroides spp. in environmental waters,12,29 and other nonhuman sources of E. coli such as domesticated animals. The human-specific Bacteroidales and E. coli uidA gene marker showed moderate correlation (r2 = 0.17, p-value < 0.0001) (Figure 3), indicating that the water was mainly contaminated

determine the amplification efficiencies and linear range of the real-time PCR assays. The qPCR amplification efficiencies for human-specific Bacteroides 16S rRNA genetic marker, humanspecific mtDNA, and E. coli uidA gene assay ranged from 90 to 120%. Linear regression correlation coefficients (r2) were between 0.9 and 1.0 for all standard curves. The linear range of detection for qPCR assay of human-specific Bacteroides 16S rRNA genetic marker was between 102 and 106 copies, while linear range for qPCR assay of human-specific mtDNA and E. coli uidA gene were between 101 and 106 copies. Lower levels of the human-specific Bacteroides marker (

Correlative Assessment of Fecal Indicators using Human

Recommend Documents