Article pubs.acs.org/est
Detection of Fecal Bacteria and Source Tracking Identifiers in Environmental Waters Using rRNA-Based RT-qPCR and rDNA-Based qPCR Assays Tarja Pitkan̈ en, †,‡ Hodon Ryu, † Michael Elk, † Anna-Maria Hokajar̈ vi,‡ Sallamaari Siponen,‡ Asko Vepsal̈ aï nen,§ Pia Ras̈ an̈ en,‡ and Jorge W. Santo Domingo †,* †
U.S. Environmental Protection Agency, National Risk Management Research Laboratory, Cincinnati, Ohio 45268, United States National Institute for Health and Welfare, Water and Health Unit, Kuopio FI-70701, Finland § National Institute for Health and Welfare, Environmental Microbiology Unit, Kuopio FI-70701, Finland ‡
S Supporting Information *
ABSTRACT: In this study, we evaluated the use of RT-qPCR assays targeting rRNA gene sequences for the detection of fecal bacteria in water samples. We challenged the RT-qPCR assays against RNA extracted from sewage effluent (n = 14), surface water (n = 30), and treated source water (n = 15) samples. Additionally, we applied the same assays using DNA as the qPCR template. The targeted fecal bacteria were present in most of the samples tested, although in several cases, the detection frequency increased when RNA was used as the template. For example, the majority of samples that tested positive for E. coli and Campylobacter spp. in surface waters, and for human-specific Bacteroidales, E. coli, and Enterococcus spp. in treated source waters were only detected when rRNA was used as the original template. The difference in detection frequency using rRNA or rDNA (rRNA gene) was sampleand assay-dependent, suggesting that the abundance of active and nonactive populations differed between samples. Statistical analyses for each population exhibiting multiple quantifiable results showed that the rRNA copy numbers were significantly higher than the rDNA counterparts (p < 0.05). Moreover, the detection frequency of rRNA-based assays were in better agreement with the culturebased results of E. coli, intestinal enterococci, and thermotolerant Campylobacter spp. in surface waters than that of rDNA-based assays, suggesting that rRNA signals were associated to active bacterial populations. Our data show that using rRNA-based approaches significantly increases detection sensitivity for common fecal bacteria in environmental waters. These findings have important implications for microbial water quality monitoring and public health risk assessments.
1. INTRODUCTION
To circumvent the aforementioned limitations, rRNA transcripts could be used as the primary PCR-targets in environmental studies. Using rRNA transcripts is advantageous since they are the most abundant constituent of nucleic acids in bacteria16 and universal primers can be used to simultaneously study phylogenetically divergent microbial groups.17 Additionally, rRNA may be used to monitor the activity status of bacterial cells,18 and rRNA transcription rate is growth-rate dependent.19 Indeed, exponentially growing cells have greater amounts of rRNA per cell than stationary-phase cultures and nonviable cells.20 Moreover, most of the rRNA is rapidly degraded when bacterial cells reach the stationary phase21 and when cell viability decreases due to the starvation.9,22 The evaluation of rRNA-targeted qPCR assays (i.e., rRNA RT-qPCR) for the detection and quantification of fecal bacteria in environmental samples is limited,23 whereas in clinical
Quantitative PCR (qPCR) assays targeting the rRNA gene (rDNA) have been used in hundreds of environmental microbiology studies within the last ten years. Specifically, many qPCR assays have been developed and applied in the detection and quantification of fecal bacteria from water.1−5 However, the use of these assays in microbial water quality analysis has encountered some criticism as DNA may be associated with dead cells and may generally persist in the environment6,7 resulting in the overestimation of in situ active/ alive fecal bacteria.8 Due to its poor stability outside of the cell9 and its correlation with cell physiology,10 RNA has been proposed as an alternate target in the detection of environmental bacteria.11,12 Efforts targeting mRNA transcripts have been limited to studies tracking particular bacterial functional groups, and in most cases are not suited to study all members of the bacterial community.13,14 Moreover, the detection frequencies of assays targeting mRNA transcripts are often low and therefore, only abundant bacteria with high metabolic rates may be detected.15 © XXXX American Chemical Society
Received: August 7, 2013 Revised: October 30, 2013 Accepted: November 4, 2013
A
dx.doi.org/10.1021/es403489b | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
spp.37−39 present in the water samples using approved procedures. Details for each method are described in the SI. 2.3. RNA and DNA Extraction and Processing. Frozen membranes were transferred into screw-cap microcentrifuge tubes containing 40 mg of 150−600 μm diameter acid-washed DNase and RNase free glass beads (Mo Bio Laboratories, Inc., Carlsbad, CA). Cell lysates were prepared by bead-beating the membranes in 500 μL of lysis buffer (Buffer RLT Plus, Qiagen GmbH, Hilden, Germany) containing β-mercaptoethanol (Sigma-Aldrich Co., St. Louis, MO) for 40 s at maximum speed (Mini-Bead-Beater, Biospec Products, Inc., Bartlesville, OK). Membranes were centrifuged at 21 130g for 3 min and the supernatant was then transferred to spin columns from the nucleic acid extraction kit. Total RNA and DNA were extracted simultaneously from the supernatant using the AllPrep DNA/ RNA Mini Kit, following the manufacturer’s instructions (Qiagen GmbH). The elution volumes were 30 μL for RNA and 100 μL for DNA. Negative extraction controls without a membrane were also processed. RNA was further purified using Ambion TURBO DNA-free DNase kit prior to the reverse transcription step, following the manufacturer’s instructions (Life Technologies, Grand Island, NY). The RNA and DNA concentrations were measured using Qubit RNA and dsDNA HS assay kits and the Qubit 2.0 Fluorometer (Life Technologies). For RNA-based signals we used a two-step RT-qPCR approach. First, cDNA was synthesized on the extraction day from the purified total RNA extracts using random hexamer primed Superscript III system for RT-PCR, following the manufacturer’s instructions (Life Technologies). The reversetranscription was performed using 8 μL of total RNA in duplicate, producing a total of 42 μL of cDNA. Total RNA was stored at −75 °C, while cDNA and DNA extracts were stored at −20 °C until used in qPCR assays as described below. 2.4. Quantification of qPCR Marker Levels. The concentration of eight different fecal bacterial markers in water samples was measured using TaqMan qPCR assays (SI Table S2) and cDNA and DNA extracts as templates. The targeted fecal bacterial groups were Bacteroidales spp. (GenBac3 assay),40 human-specific Bacteroidales (HF183 and BacHum assays),3,41 E. coli (EC23S857 assay),42 Enterococcus spp. (Entero1 assay),43 Ent. faecalis (Faecalis assay),44 Ent. faecium (Faecium1 assay),5 and Campylobacter spp. (Camp2 assay).45 The qPCR assays were performed as previously described by Ryu et al.,5 with the exception of the Camp2 assay for which we used an annealing temperature of 58 °C (SI Table S2). Standard curves were generated using plasmids containing the sequences for each of the targeted genes.5 Undiluted, 10- and 50-fold-diluted cDNA and DNA samples were used as qPCR templates to detect the PCR inhibition.46 For samples in which PCR inhibition was detected, qPCR data was generated using the results from diluted samples (SI Table S3). Background signals detected in filtration blanks (SI Table S4) were subtracted from all the results to generate final qPCR and RT-qPCR data per assay. In most cases, the limit of detection (LOD) was set as 3 copies per reaction as suggested by Bustin et al.47 (SI Table S4). In other instances, background signals of nucleic acid extraction blanks and no-template controls were used to establish LODs. No-reverse transcription controls (undiluted and 10-fold diluted RNA samples) were used to confirm the absence of DNA in RNA extracts. 2.5. Data Analysis and Statistics. The copy number per 100 mL of water was calculated for those samples with values
studies rRNA RT-qPCR assays have been proposed for the detection of subdominant bacterial populations in human stool, blood, and skin samples.24−29 Those clinical studies have indicated that rRNA-targeted RT-qPCR assays can be more sensitive than rDNA-based qPCR assays. For example, Matsuda et al.30 reported rRNA RT-qPCR assay being 400 times more sensitive than the rDNA qPCR assay at detecting 10 different Clostridium dif f icile strains. On the basis of these studies, we hypothesize that the use of RNA as a template in RT-qPCR assays may enhance the detection frequency of fecal bacteria and the identification of fecal pollution sources in environmental waters. In this study, the detection frequency of Bacteroidales spp., human-specific Bacteroidales, Escherichia coli, Enterococcus spp., E. faecalis, E. faecium, and Campylobacter spp. in environmental waters were determined using RT-qPCR and qPCR assays. The templates used in this study were RNA and DNA extracted from sewage effluent samples, surface water samples, and treated source water samples. The primary objective of this study was to compare the detection frequency of the different rRNA gene-targeting assays using as templates both RNA and DNA. Additionally, we studied the correlation of E. coli, Enterococcus spp., and Campylobacter spp. detections by RTqPCR, qPCR, and culturable data.
2. MATERIALS AND METHODS 2.1. Description of Study Sites and Water Sampling. Water samples were collected in Finland at 10 sewage treatment plants, 15 freshwater sites and four treated source water sites from September 2011 to November 2012 (Supporting Information, SI, Figure S1, n = 59). The freshwater sites have been monitored for microbiological (including enteric pathogens) and chemical water quality parameters for over 50 years.31−33 Most samples were collected within the Kokemäki River watershed. This river is used as the primary source of water for Finland’s largest drinking water infiltration supply (Turku Region Water Ltd.). The drinking water treatment process is described in the SI. The river is also used for recreational activities (e.g., bathing, fishing). Potential fecal pollution sources are associated with urban, industrial and farming activities. The sample collected within the Loimi River watershed (which discharges to the Kokemäki River) is primarily impacted by cattle farming and other agricultural activities. Water samples were transported in coolers overnight for microbiological analyses at the National Institute for Health and Welfare, Kuopio, Finland and processed within 24 or 48 h from sample collection for culture-based methods and molecular techniques, respectively. For the molecular techniques, the samples were filtered onto 0.2 μm nylon membranes (10 mL of effluents, 300−500 mL of surface water and 1000 mL of treated source water) (N66, Ultipor, Pall Corporation, Ann Arbor, Michigan). Approximately half of the membranes (i.e., 54%) were treated with RNAlater (Qiagen, Hilden, Germany) and kept at 4 °C overnight before freezing, while the rest of the membranes were frozen immediately after filtration (SI Table S1). The membranes were shipped on dry ice to the U.S. Environmental Protection Agency, Cincinnati, OH. Autoclaved distilled water was used as negative filtration control, and the filtration blanks were prepared using the same funnels and membranes as used for filtering the samples. 2.2. Culture-Based Analyses. Culturable methods were used to enumerate E. coli,34 enterococci,35,36 and Campylobacter B
dx.doi.org/10.1021/es403489b | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Table 1. Detection Rates of the Eight TaqMan RT-qPCR and qPCR Assays sewage effluent (n = 14)
surface water (n = 30)
rDNAa general Bacteroidales (GenBac3)
rRNA
human-spesific Bacteroidales (HF183)
rRNA
human-spesific Bacteroidales (BacHum)
rRNA
E. coli (EC23S857)
rRNA
general Enterococcus (Entero1)
rRNA
Enterococcus faecalis (Faecalis)
rRNA
Enterococcus faecium (Faecium1)
rRNA
Campylobacter spp. (Camp2)
rRNA
− + − + − + − + − + − + − + − +
treated source water (n = 15)
rDNA
−
+
pb
0 0 1 0 1 0 2 1 0 1 3 6 3 4 2 2
0 14 1 12 0 13 0 11 0 13 0 5 0 7 0 10
NA 1.00 1.00 1.00 1.00 0.031 0.125 0.500
− + − + − + − + − + − + − + − +
rDNA
−
+
p
0 0 4 1 2 0 10 15 0 8 27 3 28 2 4 13
2 28 6 19 3 25 1 4 1 21 0 0 0 0 1 12
0.500
− + − + − + − + − + − + − + − +
0.125 0.250 0.001 0.039 0.250 0.500 0.002
−
+
p
8 0 10 5 8 5 9 4 5 10 15 0 15 0 15 0
3 4 0 0 1 1 2 0 0 0 0 0 0 0 0 0
0.250 0.063 0.219 0.688 0.002 NA NA NA
Number of samples; − = samples with results below the limit of detection, + = samples with results above the limit of detection. bNonparametric related samples McNemar test to compare the difference between the detection rates of rRNA and rDNA targeted assays. p < 0.05 indicate statistically significant difference (in bold). NA = not applicable.
a
Table 2. Number of Samples with Results above Limit of Quantification and the Median (Min−Max) Concentrations of the Markers (log10 GC 100 mL−1) in TaqMan RT-qPCR and qPCR Assays sewage effluent (n = 14) assay general Bacteroidales (GenBac3) human-spesific Bacteroidales (HF183) human-spesific Bacteroidales (BacHum) E. coli (EC23S857) general Enterococcus (Entero1) Enterococcus faecalis (Faecalis) Enterococcus faecium (Faecium1) Campylobacter spp. (Camp2)
surface water (n = 30)
treated source water (n = 15)
target
N
median
min−max
pa
N
median
min−max
pa
N
median
min−max
pa
rRNA rDNA rRNA rDNA rRNA rDNA rRNA rDNA rRNA rDNA rRNA rDNA rRNA rDNA rRNA rDNA
14 14 12 12 13 13 11 11 14 13 8 1 6 0 12 7
9.09 7.11 7.17 5.75 7.84 6.35 5.69 4.04 7.70 5.70 4.96 3.74 4.81 − 4.71 3.64
5.99−9.88 4.43−8.08 4.59−7.75 3.29−6.46 3.74−8.75 2.72−7.19 5.09−6.42 3.40−4.49 4.19−8.55 3.93−6.64 4.49−5.48 3.74 4.34−5.59 − 3.40−5.52 3.17−3.88
0.001
28 30 20 21 24 24 14 2 24 12 1 0 0 0 19 7
5.57 4.33 3.48 2.66 4.07 2.98 2.86 2.20 4.82 3.33 3.45 − − − 3.09 2.38
3.84−6.27 3.32−4.90 2.62−4.62 2.29−3.41 2.97−5.11 2.06−4.01 2.36−3.56 2.20−2.21 3.53−5.80 3.03−4.06 3.45 − − − 2.49−3.98 2.25−2.90
