w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Opportunistic pathogens in roof-captured rainwater samples, determined using quantitative PCR W. Ahmed a,*, H. Brandes a,b, P. Gyawali a,c, J.P.S. Sidhu a, S. Toze a,c a
CSIRO Land and Water, Ecosciences Precinct, 41 Boggo Road, Brisbane 4102, Australia Environmental Studies Department, University of Colorado, Boulder, CO 80309-0397, USA c School of Population Health, University of Queensland, Herston Road, Brisbane 4006, Australia b
article info
abstract
Article history:
In this study, quantitative PCR (qPCR) was used for the detection of four opportunistic
Received 16 September 2013
bacterial pathogens in water samples collected from 72 rainwater tanks in Southeast
Received in revised form
Queensland, Australia. Tank water samples were also tested for fecal indicator bacteria
5 December 2013
(Escherichia coli and Enterococcus spp.) using culture-based methods. Among the 72 tank
Accepted 13 December 2013
water samples tested, 74% and 94% samples contained E. coli and Enterococcus spp.,
Available online 9 January 2014
respectively, and the numbers of E. coli and Enterococcus spp. in tank water samples ranged
Keywords:
and 6% of tank water samples contained Aeromonas hydrophila, Staphylococcus aureus,
from 0.3 to 3.7 log10 colony forming units (CFU) per 100 mL of water. In all, 29%, 15%, 13%, Roof-captured rainwater
Pseudomonas aeruginosa and Legionella pneumophila, respectively. The genomic units (GU) of
Fecal indicator bacteria
opportunistic pathogens in tank water samples ranged from 1.5 to 4.6 log10 GU per 100 mL
Opportunistic pathogens
of water. A significant correlation was found between E. coli and Enterococcus spp. numbers
Quantitative PCR
in pooled tank water samples data (Spearman’s rs ¼ 0.50; P < 0.001). In contrast, fecal in-
Health risks
dicator bacteria numbers did not correlate with the presence/absence of opportunistic pathogens tested in this study. Based on the results of this study, it would be prudent, to undertake a Quantitative Microbial Risk Assessment (QMRA) analysis of opportunistic pathogens to determine associated health risks for potable and nonpotable uses of tank water. Crown Copyright ª 2014 Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Roof-captured rainwater has been used as potable and nonpotable water sources in many countries (Despins et al., 2009; Evans et al., 2006; Uba and Aghogho, 2000). There are several advantages of using roof-captured rainwater, including (i) reducing the pressure on the mains water supply, (ii) providing an alternative water supply during times of water restrictions,
and (iii) reducing stormwater runoff that can often degrade creek ecosystem health. Despite these advantages, roofcaptured rainwater has not been widely utilized for potable purpose due to lack of information on the risk from exposure to pathogenic bacteria and protozoa. The presence of enteric pathogens such as Campylobacter spp., Salmonella spp., Giardia spp., and Cryptosporidium spp. in tank water samples has been reported (Albrechtsen, 2002; Crabtree et al., 1996; Savill et al., 2001; Simmons et al., 2001). Case control studies have also
* Corresponding author. Tel.: þ61 7 3833 5582; fax: þ61 7 3833 5503. E-mail addresses: [email protected], [email protected] (W. Ahmed). 0043-1354/$ e see front matter Crown Copyright ª 2014 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2013.12.021
362
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
established links between gastroenteritis and consumption of untreated tank water (Brodribb et al., 1995; Merritt et al., 1999). In contrast, there is a general community perception that tank water is safe to drink without having to undergo prior treatment. This was further supported by the epidemiological surveys that suggested tank water does not pose risk of gastroenteritis (Heyworth et al., 2006; Rodrigo et al., 2010). However, such results should be interpreted with care due to the lack of sensitivity of the epidemiological tool to detect gastroenteritis (Hrudey and Hrudey, 2004). In addition, considering the high costs and time required, epidemiological studies may not be practical for the sensitive detection of impacts on gastroenteritis in the community. Heyworth et al. (2006) also pointed out a level of acquired immunity among roof-captured rainwater users, and therefore, may not reflect the actual risk to new users. The cases of gastroenteritis or other infections due to potable use of tank water could actually be masked by the background levels of gastroenteritis or other infections from other sources, such as consumption of food and community-based infections. Therefore, legitimate questions have arisen regarding the microbiological quality of tank water and consequent public health risks. The microbiological quality of tank water is generally assessed by monitoring fecal indicator bacteria such as Escherichia coli and Enterococcus spp. The most important limitation of fecal indicator bacteria arises from their poor correlation with pathogens in water resources (Ahmed et al., 2009; Ho¨rman et al., 2004; McQuaig et al., 2006). This is not unexpected, considering the fact that fecal indicator bacteria exhibit different survival rates compared to pathogens, especially viruses and protozoa. Currently, there is a paucity of knowledge in relation to the occurrence and quantitative numbers of pathogens in tank water and their relationships with fecal indicator bacteria. To measure health risks, microbial assessment should involve the analysis of tank water samples for actual pathogens, not just the fecal indicator bacteria. Limited information, however, is available on the occurrence of pathogens especially opportunistic pathogens in tank water samples. An opportunistic pathogen is defined as one that usually does not cause diseases in healthy individuals, however, may cause diseases when the immune systems of hosts are compromised. There has been an increased interest in opportunistic pathogens with the increase in the immunocompromised population (Schoen and Ashbolt, 2011). For example, Aeromonas hydrophila is associated with both diarrheal and extraintestinal infections in humans (Altwegg and Geiss, 1989). Immunocompromised patients can develop sepsis or meningitis, and both healthy humans and immunocompromised patients can suffer from Aeromonas wound infections. Other opportunistic pathogens include Pseudomonas aeurginosa, Legionella pneumophila and Staphylococcus aureus. P. aeurginosa is a major cause of hospital-acquired infections with a high mortality rate (Rusin et al., 1997). L. pneumophila is a major cause of an estimated 8000e18,000 reported cases of legionellosis with mortality rate around 9% in the United States every year (CDC, 2011; Marston et al., 1997). S. aureus is the most common cause of bloodstream mortality and morbidity in nosocomial and community settings (Bassetti et al., 2012).
Various opportunistic bacterial pathogens such as Pseudomonas spp., Aeromonas spp., and Legionella spp. have been isolated from drinking water sources (Baker and Hegarty, 2001). Additional testing of tank water samples is, therefore, required to obtain quantitative information on these opportunistic pathogens so that information can be given to the public health regulators who are charged with protecting public health. The main aim of this pilot study was to determine the frequency of occurrence and numbers of four opportunistic pathogens (A. hydrophila, L. pneumophila, P. aeruginosa, and S. aureus) in tank water samples, and whether fecal indicator bacteria monitoring might reflect the presence/absence of these opportunistic pathogens.
2..
Materials and methods
2.1.
Tank water sampling
In all, 72 rainwater tank samples were collected from 72 residential houses representing 18 suburbs in Brisbane and Gold Coast region in Southeast Queensland, Australia in MayeJuly 2012. The samples were collected within one to four days after rainfall events. The size of the tanks sampled ranged from 1000 to 30,000 L, and the end uses were (i) potable use (58%), and (ii) nonpotable use (42%). Water samples were collected in sterilized 20-L containers from the outlet taps located close to the base of the tanks. Before the tank was sampled, the tap was allowed to run for 30e60 s to flush out water from the tap. Samples were transported to the laboratory and processed within 2e4 h.
2.2.
Enumeration of fecal indicator bacteria
E. coli and Enterococcus spp. numbers in tank water samples were enumerated using membrane filtration method (US EPA, 1997; US EPA, 2002). Sample serial dilutions were made and filtered through 0.45 mm pore sized (47 mm diameter) nitrocellulose membranes (Millipore, Tokyo, Japan). The membranes were placed on modified mTEC agar (Difco, Detroit, MI) and membrane-Enterococcus indoxyl-b-D-glucoside (mEI) agar (Difco) for the isolation of E. coli and Enterococcus spp., respectively. Modified mTEC agar plates were incubated at 35 C for 2 h to recover stressed cells, followed by incubation at 44 C for 22 h (US EPA, 2002), and mEI agar plates were incubated at 41 C for 48 h (US EPA, 1997).
2.3.
Concentration of tank water samples
Approximately, 19 L water sample from each rainwater tank was concentrated by hollow-fiber ultrafiltration system (HFUS), using Hemoflow HF80S dialysis filters (Fresenius Medical Care, Lexington, MA, USA) as previously described (Hill et al., 2005). Each sample was concentrated to approximately 100 mL. Each 100 mL sample was further centrifuged at 3000 g for 30 min at 4 C. The supernatant was discarded, and the pellet was resuspended in 5 mL of sterile distilled water.
363
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Table 1 e - Primer sequences for quantitative PCR (qPCR) assays for the quantitative detection of opportunistic pathogens. Pathogens
Primer sequence (50 e30 )a
A. hydrophila
F e AAC CTG GTT CCG CTC AAG CCG TTG R e TTG CTC GCC TCG GCC CAG CAG CT F e GCA ATG TCA ACA GCAA R e CAT AGC GTC TTG CATG F e TGC TGG TGG CAC AGG ACA T R e TTG TTG GTG CAG TTC CTC ATT G F e CGT ATT AGC AGA GAG CCA ACC A R e GTG ATT TTA CTC GCT TTG TGC AA
L. pneumophila P. aeruginosa S. aureus a
Genes
Functions
References
lip
Novel lipase
Casco´n et al., 1996
mip
Surface protein virulence factor
Wilson et al., 2003
regA
Toxin A synthesis regulating gene
Shannon et al., 2007
sec
Enterotoxin C gene
Shannon et al., 2007
F, forward primer; R, reverse primer.
2.4.
DNA extraction
For quantitative PCR (qPCR) analysis, DNA was extracted from the pellet obtained from 1.5 mL (equivalent to 5.7 L of water) of concentrated samples using a DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA). All DNA samples were quantified using a NanoDrop spectrophotometer (ND-1000, NanoDrop Technology, Wilmington, DE). Each DNA sample was amplified using a universal bacterial PCR assay to confirm successful DNA extraction process (Boon et al., 2000).
2.5.
Preparation of qPCR standard curves
To obtain qPCR standard curves, DNA was extracted from the 1 mL of broth cultures of A. hydrophila ATCC 7966, L. pneumophila ATCC 33152, P. aeruginosa ATCC 15629, and S. aureus ATCC 12600 using DNeasy Blood and Tissue Kit (Qiagen). The concentrations of genomic DNA samples were determined using a NanoDrop spectrophotometer (ND-1000, NanoDrop Technology). The gene copies were calculated, and a 10-fold dilution ranging from 106 to 101 copies per mL of DNA extract was prepared from the genomic DNA, and stored at 20 C until use. For each standard, the genomic copies were plotted against the cycle number at which the fluorescence signal increased above the threshold value (CT value).
2.6.
qPCR primers and conditions
The primers used for qPCR assays are shown in Table 1. qPCR amplifications were performed in 20-mL reaction mixtures using Sso Fast EvaGreen Supermix (Bio-Rad Laboratories, Calif). The PCR mixture contained 10 mL of Supermix, 300e500 nM each primer, 5 mL of template DNA and DNaseand RNase-free deionized water. The qPCR conditions were as follows: 15 min at 95 C, 45 cycles of 60 s at 94 C, 60 s at 62 C, and 90 s at 72 C (for A. hydrophila), 10 min at 95 C, 45 cycles of 15 s at 95 C, 60 s at 60 C (for P. aeurginosa and S. aureus), and 15 min at 95 C, 45 cycles of 30 s at 94 C, 60 s at 54 C, and 60 s at 72 C (for L. pneumophila). For each qPCR experiment, a negative control (sterile water) was included. The qPCR assays were performed using the Bio-Rad iQ5 (Bio-Rad Laboratories). To separate the specific product from non-specific products, DNA melting curve analysis was performed for each qPCR assay.
2.7. qPCR assays reproducibility and lower limit of detection (LLOD) The reproducibility of the qPCR assays was assessed by determining intra-assay repeatability and inter-assay reproducibility. The coefficient of variation (CV) was calculated using six dilutions (106 to 101 gene copies) of genomic DNA isolated from opportunistic pathogens. Each dilution was quantified in replicates. The CV for evaluation of intra-assay repeatability was calculated based on the CT value by testing the six dilutions six times in the same experiment. The CV for inter-assay reproducibility was calculated based on the CT values of six dilutions on three different days. To determine qPCR LLOD, extracted genomic DNA samples of opportunistic pathogens were quantified using a NanoDrop spectrophotometer (ND-1000, NanoDrop Technology). Ten-fold serial dilutions were made and tested with the qPCR assays. The lowest amount of DNA detected in replicate assays was considered qPCR LLOD.
2.8.
Recovery efficiency
To determine the method recovery efficiency, E. coli was used as a representative microorganism. The recovery efficiency of E. coli was assumed to be similar to those opportunistic pathogens tested in this study. Nineteen liters deionized water (n ¼ 3) and 19 L tank water (n ¼ 3) samples were spiked with known cells of E. coli. In brief, E. coli (ACM 1803) was grown overnight in Nutrient Broth (Oxoid, UK), and the number of cells in the broth culture was determined using a spread plate method. Ten-fold serial dilutions (2.1 106, 2.1 105, 2.1 104) were made and seeded into the distilled water and tank water samples. The samples were concentrated, and DNA extraction was performed according to the method described above. All DNA samples were tested in triplicate with qPCR of E. coli 23S rRNA gene (Chern et al., 2011). The recovery efficiency (%) was calculated using the following equation: recovery (%) ¼ [Genomic units (GU) of E. coli 23S rRNA after qPCR analysis/seeded GU of E. coli 23S rRNA before concentration process] 100.
2.9.
PCR inhibitors
An experiment was conducted to determine the potential presence of PCR inhibitors in tank water samples collected randomly from 10 different tanks. Ten-fold serial dilutions
364
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Table 2 e The amplification efficiency, correlation coefficient (r2), intra-assay and inter-assay coefficient of variation (CV) for the qPCR assays within the range of 106 to 102 GU of opportunistic pathogens. Opportunistic pathogens
Amplification efficiency (%)
r2
98 106 98 103
0.99 0.98 0.99 0.95
Mean SD of CV Intra-assay
A. hydrophila L. pneumophila P. aeruginosa S. aureus
0.86 0.84 0.94 0.82
0.1 0.1 0.1 0.2
Inter-assay 3.1 4.3 3.6 2.9
0.7 0.6 1.0 0.9
SD: Standard deviation.
were made for each sample. All DNA samples (undiluted, 10fold diluted, and 100-fold diluted) were seeded with 103 gene copies of human specific HF183 Bacteroides markers (Seurinck et al., 2005). The CT values obtained for the HF183 spiked DNA samples were compared to those of the distilled water to obtain information on the level of PCR inhibitors.
2.10.
Statistical analysis
occurring at lower cell counts (seeded). The mean recovery efficiency was 72% 9%. The estimated recovery efficiency in tank water samples ranged between 55% and 71% with the greatest variability occurring at lower cell counts (seeded). The mean recovery efficiency was 60% 8%.
3.3.
PCR inhibitors
Fecal indicator bacteria and opportunistic pathogen data were log10 transformed. The nonparametric Spearman rank correlation was performed to obtain correlation between CFU of E. coli and Enterococcus spp. A binary logistic regression analysis was also performed to obtain correlations between the presence/absence results of opportunistic pathogens and the CFU of E. coli and Enterococcus spp. A difference was considered significant if the P value for the model chi-square was 0.05 and the confidence interval for the odds ratio did not overlap. IBM SPSS software (version 21.0) was used for all statistical analysis.
The mean CT value for the HF183 spiked distilled water DNA was 23.8 0.4. For tank water samples, the mean CT value was 22.8 0.5 when undiluted DNA was spiked with the HF183. For 10-fold and 100-fold dilutions of DNA spiked with the HF183, these values were 23.2 0.6 and 23.8 0.3, respectively. Analysis of variance (ANOVA) was performed to determine the differences between the CT values obtained for distilled water and those obtained for tank water samples. No significant differences were observed between the CT values for the HF183 spiked distilled water, undiluted DNA, and serially diluted DNA (10-fold and 100-fold), thus indicating that the tested tank water samples were free of PCR inhibitors.
3.
Results
3.4.
3.1.
qPCR standards, assay reproducibility and LLOD
Culturable E. coli exceeded that specified by Australian Drinking Water Guidelines for 74% of 72 tank water samples. The Colony forming units (CFU) of E. coli in positive samples ranged between 0.5 and 3.7 log10 CFU per 100 mL of water with an average of 1.8 log10 CFUs per 100 mL (Fig. 1). Similarly, 94% samples were also positive for Enterococcus spp., and the
qPCR standards for target opportunistic pathogens were analyzed in order to determine the reaction efficiencies. The standard curves had a linear range of quantification from 106 to 101 gene copies per mL of DNA extracts. The amplification efficiencies ranged from 98% to 106%, and the correlation coefficient (r2) ranged from 0.95 to 0.99 for all qPCR assays. The mean intra-assay and inter-assay coefficient of variation (CV) values and standard deviations (SD) were less than 1% and 5%, respectively, indicating high reproducibility (Table 2). LLOD assays were performed using purified genomic DNA isolated from pure culture of each opportunistic pathogen. To determine the reproducibility of the assays, three replicates were tested for each target pathogen. The real-time PCR LLOD ranged from 2 to 8 fg for L. pneumophila, P. aeruginosa and S. aureus. For A. hydrophila, the LLOD was 13 fg.
3.2.
E. coli and Enterococcus spp. numbers
Recovery efficiency
The recovery efficiency was determined by seeding tank water and distilled water samples with known numbers of E. coli. The estimated recovery efficiency in distilled water samples ranged between 65% and 85% with the greatest variability
Fig. 1 e Box-and-whisker plots of the log10 CFU of fecal indicator bacteria in tank water samples. The inner box lines represent the medians, while the outer box lines represent the 25th and 75th percentiles.
365
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Table 3 e Percentage of tank water samples that were positive and quantifiable for opportunistic pathogens tested in this study. Opportunistic pathogens A. hydrophila L. pneumophila P. aeruginosa S. aureus a b
No. of positive samples/no. of total samples tested (% of positive samples)a
No. of quantifiable samples/no. of total samples tested (% of quantifiable samples)b
21/72 (29) 5/72 (6) 9/72 (13) 11/72 (15)
14/72 (19) 5/72 (6) 4/72 (6) 10/72 (14)
: Real-time PCR. : qPCR.
numbers ranged between 0.3 and 3.6 log10 CFUs with an average of 1.5 log10 CFU per 100 mL. Among the 72 tank water samples tested, 72% were positive for both E. coli and Enterococcus spp., and 97% were positive for either E. coli or Enterococcus spp. numbers. A nonparametric Spearman rank correlation was performed to determine whether any correlation exists between E. coli and Enterococcus spp. in tank water samples. Significant correlation was found between these two fecal indicator bacteria (Spearman’s rs ¼ 0.50; P < 0.001).
S. aureus (2.8e4.6 log10 GU with an average 3.8 log10 GU) per 100 mL of water (Fig. 2). Binary logistic regression analysis was used to identify whether any correlation existed between the numbers of E. coli or Enterococcus spp. and the presence/ absence results for the opportunistic pathogens (Table 4). The presence/absence results of the opportunistic pathogens did not correlate with E. coli and Enterococcus spp. numbers.
4. 3.5. Genomic units (GU) of opportunistic pathogens in tank water samples Of the 72 tank water samples tested, A. hydrophila was the most prevalent (29%) among all the opportunistic pathogens tested in this study (Table 3). The prevalence of other opportunistic pathogens was as follows: 15% samples were positive for S. aureus, 13% were positive for P. aeruginosa. The prevalence of L. pneumophila (6%) was low compared to other opportunistic pathogens. Of the 72 tank samples, 8% were positive for two opportunistic pathogens, and 32% were positive for at least one opportunistic pathogen. In contrast, none of these opportunistic pathogens was detected in 56% tank water samples. A number of samples although gave PCR positive signal but was not quantifiable hence, not included in the qPCR analysis. The range of GU was as follows: A. hydrophila (2.1 and 4.5 log10 with an average 3.5 log10 GU), L. pneumophila (3.2e4 log10 GU with an average 3.4 log10 GU), P. aeurginosa (1.5e4.3 log10 GU with an average 3.4 log10 GU) and
Discussion
In this study, roof-captured rainwater samples were tested for the numbers of E. coli and Enterococcus spp., using conventional culture-based methods. In addition, qPCR assays were used to obtain information on the GU of four opportunistic bacterial pathogens. The application of PCR-based methods has generated interest in the direct monitoring of pathogens in water because they are rapid, sensitive and can detect microorganisms that are difficult to grow using conventional culture methods. In view of this, qPCR assays were used to quantify opportunistic pathogens in tank water samples. The detection sensitivity of the qPCR assays used in this study was evaluated by amplifying known concentrations of DNA from target opportunistic pathogens. The LLOD ranged from 0.2 to 13 fg for the tested opportunistic pathogens indicating that the detection sensitivity values of our qPCR assays were comparable to the values reported in the literature (Behets et al., 2007; Sails et al., 2002). Prior to setting up qPCR assays, specificity of each primer was determined by searching for similar sequences in microbial genomes using the Basic Local
Table 4 e Correlations between colony forming units (CFUs) of fecal indicator bacteria and the presence/ absence of opportunistic pathogens.a Relationship E. coli vs. A. hydrophila E. coli vs. L. pneumophila E. coli vs. P. aeurginosa E. coli vs. S. aureus Enterococcus spp. vs. A. hydrophila Enterococcus spp. vs. L. pneumophila Enterococcus spp. vs. P. aeurginosa Enterococcus spp. vs. S. aureus
Fig. 2 e - Box-and-whisker plots of the log10 GU of opportunistic pathogens in tank water samples. The inner box lines represent the medians, while the outer box lines represent the 25th and 75th percentiles.
a
R2
P value
Odd ratio
0.027 0.001 0.011 0.006 0.004 0.056 0.071 0.001
0.275 0.854 0.481 0.568 0.660 0.454 0.095 0.845
1.000 1.000 1.000 1.000 1.000 0.994 1.001 1.000
Data are results of binary logistic regression analysis (P values) performed to identify correlations between the CFU of fecal indicator bacteria and the presence/absence of opportunistic pathogens in roof-captured rainwater samples.
366
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Alignment Search Tool (BLAST) program (http://www.ncbi. nlm.nih.gov/BLAST/). The cross-reactivity of each primer set for each target has also been assessed against a panel of 20 microorganisms (data not shown). The primers did not amplify any PCR products other than those products that were expected, suggesting a high specificity for the primers used in this study. In the 72 tank water samples tested, 74% samples had >1 CFU E. coli per 100 mL of water, exceeding Australian Drinking Water Guidelines (ADWG, 2011). The numbers of E. coli and Enterococcus spp. were highly variable in the tank water samples, around 19% and 26% tank water samples had >2 log10 CFU E. coli and Enterococcus spp. per 100 mL of water, suggesting the occurrence of fecal pollution from birds, mammals insects and perhaps environmental sources (Ahmed et al., 2012a, 2012b). This was not unexpected because the samples were collected within one to four days after rainfall events, when fecal and other organic matter deposited on the roof enters tanks via roof runoff. Enterococcus spp. was more prevalent than E. coli in tank water samples. A number of samples (17 of 72) were culture positive for Enterococcus spp. but negative for E. coli, possibly because Enterococcus spp. persist longer in the tank water than E. coli (Ahmed et al., 2013). It is also possible that all Enterococcus spp. isolated from tank water samples were not from a fecal origin. A recent study reported the presence of soil or plant related Enterococcus mundtii and Enterococcus casseliflavus in a number of tank water samples in Southeast Queensland, Australia (Ahmed et al., 2012b). In this study, opportunistic pathogens were detected in one or more tank water samples. The overall prevalence of S. aureus, P. aeruginosa, and A. hydrophila (ranged from 13 to 29%) in tank water samples was higher than L. pneumophila (6%). The occurrence of L. pneumophila in this study, however, was lower than previously reported studies (Albrechtsen, 2002; Broadhead et al., 1998). The occurrence of A. hydrophila in this study is comparable to previous studies undertaken in Australia and New Zealand (Simmons et al., 2001; CRC for Water Quality Treatment, 2006). The occurrence of pathogens in tank water samples in different studies is expected to vary due to several factors such as sample volume, detection method sensitivity, climatic conditions and physical parameters of the rainwater system. Another important factor which may also influence the occurrence of pathogens is the types of wildlife that has access to roof. In Southeast Queensland, for example, marsupial possums are a common animal in and around urban dwellings and frequently traverse roof tops in their nocturnal movements. Birds and wildlife including possums and other animals, therefore, could be a major source of certain bacterial pathogens in tank water in Southeast Queensland (Ahmed et al., 2012b). In this study, one or more opportunistic pathogens were quantitatively detected in a number of nonpotable tank water samples. It can be postulated that the magnitude of health risks from nonpotable uses could be lower than potable use due to lower exposure levels (Ahmed et al., 2010). While much of the focus on bacterial pathogens in drinking water is predominantly on enteric indicators (E. coli) and pathogens (Salmonella spp.) in tank water, opportunistic pathogens tested in
this study are also considered to be of concern for human health. In this study, the high level of GU (average 3.2 to 3.8 log10 per 100 mL) of opportunistic pathogens along with E. coli and Enterococcus spp. in potable tank water samples indicate a poor level of microbial quality, and could represent potential health risks to users with low immunity. The result is particularly of significance because a recent study has demonstrated the inefficacy of the under sink filtration methods that are being used for tank water purification prior to drinking in Southeast Queensland, Australia (Ahmed et al., 2012a). There have been concerns that the presence of these opportunistic pathogens in drinking water may represent public health risks (Egorov et al., 2011; Farkas et al., 2012). High numbers of S. aureus (average 3.8 log10) were detected in certain tank water samples. S. aureus is known to cause food poisoning, a variety of skin abscesses, pustules and septicemia (LeChevallier and Seidler, 1980). A. hydrophila is considered to be of concern because of the high GU (average 3.5 log10) detected per 100 mL of water in this study. A recent study demonstrated a link between Aeromonas spp. isolated from clinical and water samples, indicating a transmission from water (Khajanchi et al., 2010). The high GU of L. pneumophila, in certain tank water samples can be considered to be of concern because direct exposure to aerosols containing Legionella spp. generated from cooling towers, spa and shower heads are implicated as the vehicle for bacterial transmission (Armstrong and Haas, 2007a; Bauer et al., 2008; Schoen and Ashbolt, 2011). We acknowledge that qPCR results do not provide information regarding the infectivity status of the target microorganisms. Another major limitation of PCR is its inability to distinguish between viable and nonviable pathogens. Therefore, over estimation of opportunistic pathogens cannot be ruled out (Krøjgaard, 2011). In contrast, several method comparison studies also reported high numbers of qPCR quantified E. coli and Enterococcus spp. cell equivalents compared to culture-based CFU in recreational waters (Converse et al., 2011; Haugland et al., 2005). A possible explanation for such discrepancy is that certain microorganism including pathogens can enter viable but not culturable (VBNC) state when subjected to adverse environmental conditions (Kolling and Matthews, 2001; Heim et al., 2002). In addition, culture-based methods may underestimate the bacterial numbers due to presence of injured or stressed cells (Juhna et al., 2007). Another important factor that needs to be recognized is that some target opportunistic pathogens tested in this study have high ID50 levels. For example, large oral doses up to 1010 CFU, A. hydrophila failed to produce diarrhea in human volunteers (Morgan et al., 1985). Therefore, the level of A. hydrophila GU that were detected in tank water samples may not pose any risks to healthy humans, however, immunocompromised and elderly population could be at risk. For L. pneumophila, a small dose such as
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Opportunistic pathogens in roof-captured rainwater samples, determined using quantitative PCR W. Ahmed a,*, H. Brandes a,b, P. Gyawali a,c, J.P.S. Sidhu a, S. Toze a,c a
CSIRO Land and Water, Ecosciences Precinct, 41 Boggo Road, Brisbane 4102, Australia Environmental Studies Department, University of Colorado, Boulder, CO 80309-0397, USA c School of Population Health, University of Queensland, Herston Road, Brisbane 4006, Australia b
article info
abstract
Article history:
In this study, quantitative PCR (qPCR) was used for the detection of four opportunistic
Received 16 September 2013
bacterial pathogens in water samples collected from 72 rainwater tanks in Southeast
Received in revised form
Queensland, Australia. Tank water samples were also tested for fecal indicator bacteria
5 December 2013
(Escherichia coli and Enterococcus spp.) using culture-based methods. Among the 72 tank
Accepted 13 December 2013
water samples tested, 74% and 94% samples contained E. coli and Enterococcus spp.,
Available online 9 January 2014
respectively, and the numbers of E. coli and Enterococcus spp. in tank water samples ranged
Keywords:
and 6% of tank water samples contained Aeromonas hydrophila, Staphylococcus aureus,
from 0.3 to 3.7 log10 colony forming units (CFU) per 100 mL of water. In all, 29%, 15%, 13%, Roof-captured rainwater
Pseudomonas aeruginosa and Legionella pneumophila, respectively. The genomic units (GU) of
Fecal indicator bacteria
opportunistic pathogens in tank water samples ranged from 1.5 to 4.6 log10 GU per 100 mL
Opportunistic pathogens
of water. A significant correlation was found between E. coli and Enterococcus spp. numbers
Quantitative PCR
in pooled tank water samples data (Spearman’s rs ¼ 0.50; P < 0.001). In contrast, fecal in-
Health risks
dicator bacteria numbers did not correlate with the presence/absence of opportunistic pathogens tested in this study. Based on the results of this study, it would be prudent, to undertake a Quantitative Microbial Risk Assessment (QMRA) analysis of opportunistic pathogens to determine associated health risks for potable and nonpotable uses of tank water. Crown Copyright ª 2014 Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Roof-captured rainwater has been used as potable and nonpotable water sources in many countries (Despins et al., 2009; Evans et al., 2006; Uba and Aghogho, 2000). There are several advantages of using roof-captured rainwater, including (i) reducing the pressure on the mains water supply, (ii) providing an alternative water supply during times of water restrictions,
and (iii) reducing stormwater runoff that can often degrade creek ecosystem health. Despite these advantages, roofcaptured rainwater has not been widely utilized for potable purpose due to lack of information on the risk from exposure to pathogenic bacteria and protozoa. The presence of enteric pathogens such as Campylobacter spp., Salmonella spp., Giardia spp., and Cryptosporidium spp. in tank water samples has been reported (Albrechtsen, 2002; Crabtree et al., 1996; Savill et al., 2001; Simmons et al., 2001). Case control studies have also
* Corresponding author. Tel.: þ61 7 3833 5582; fax: þ61 7 3833 5503. E-mail addresses: [email protected], [email protected] (W. Ahmed). 0043-1354/$ e see front matter Crown Copyright ª 2014 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2013.12.021
362
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
established links between gastroenteritis and consumption of untreated tank water (Brodribb et al., 1995; Merritt et al., 1999). In contrast, there is a general community perception that tank water is safe to drink without having to undergo prior treatment. This was further supported by the epidemiological surveys that suggested tank water does not pose risk of gastroenteritis (Heyworth et al., 2006; Rodrigo et al., 2010). However, such results should be interpreted with care due to the lack of sensitivity of the epidemiological tool to detect gastroenteritis (Hrudey and Hrudey, 2004). In addition, considering the high costs and time required, epidemiological studies may not be practical for the sensitive detection of impacts on gastroenteritis in the community. Heyworth et al. (2006) also pointed out a level of acquired immunity among roof-captured rainwater users, and therefore, may not reflect the actual risk to new users. The cases of gastroenteritis or other infections due to potable use of tank water could actually be masked by the background levels of gastroenteritis or other infections from other sources, such as consumption of food and community-based infections. Therefore, legitimate questions have arisen regarding the microbiological quality of tank water and consequent public health risks. The microbiological quality of tank water is generally assessed by monitoring fecal indicator bacteria such as Escherichia coli and Enterococcus spp. The most important limitation of fecal indicator bacteria arises from their poor correlation with pathogens in water resources (Ahmed et al., 2009; Ho¨rman et al., 2004; McQuaig et al., 2006). This is not unexpected, considering the fact that fecal indicator bacteria exhibit different survival rates compared to pathogens, especially viruses and protozoa. Currently, there is a paucity of knowledge in relation to the occurrence and quantitative numbers of pathogens in tank water and their relationships with fecal indicator bacteria. To measure health risks, microbial assessment should involve the analysis of tank water samples for actual pathogens, not just the fecal indicator bacteria. Limited information, however, is available on the occurrence of pathogens especially opportunistic pathogens in tank water samples. An opportunistic pathogen is defined as one that usually does not cause diseases in healthy individuals, however, may cause diseases when the immune systems of hosts are compromised. There has been an increased interest in opportunistic pathogens with the increase in the immunocompromised population (Schoen and Ashbolt, 2011). For example, Aeromonas hydrophila is associated with both diarrheal and extraintestinal infections in humans (Altwegg and Geiss, 1989). Immunocompromised patients can develop sepsis or meningitis, and both healthy humans and immunocompromised patients can suffer from Aeromonas wound infections. Other opportunistic pathogens include Pseudomonas aeurginosa, Legionella pneumophila and Staphylococcus aureus. P. aeurginosa is a major cause of hospital-acquired infections with a high mortality rate (Rusin et al., 1997). L. pneumophila is a major cause of an estimated 8000e18,000 reported cases of legionellosis with mortality rate around 9% in the United States every year (CDC, 2011; Marston et al., 1997). S. aureus is the most common cause of bloodstream mortality and morbidity in nosocomial and community settings (Bassetti et al., 2012).
Various opportunistic bacterial pathogens such as Pseudomonas spp., Aeromonas spp., and Legionella spp. have been isolated from drinking water sources (Baker and Hegarty, 2001). Additional testing of tank water samples is, therefore, required to obtain quantitative information on these opportunistic pathogens so that information can be given to the public health regulators who are charged with protecting public health. The main aim of this pilot study was to determine the frequency of occurrence and numbers of four opportunistic pathogens (A. hydrophila, L. pneumophila, P. aeruginosa, and S. aureus) in tank water samples, and whether fecal indicator bacteria monitoring might reflect the presence/absence of these opportunistic pathogens.
2..
Materials and methods
2.1.
Tank water sampling
In all, 72 rainwater tank samples were collected from 72 residential houses representing 18 suburbs in Brisbane and Gold Coast region in Southeast Queensland, Australia in MayeJuly 2012. The samples were collected within one to four days after rainfall events. The size of the tanks sampled ranged from 1000 to 30,000 L, and the end uses were (i) potable use (58%), and (ii) nonpotable use (42%). Water samples were collected in sterilized 20-L containers from the outlet taps located close to the base of the tanks. Before the tank was sampled, the tap was allowed to run for 30e60 s to flush out water from the tap. Samples were transported to the laboratory and processed within 2e4 h.
2.2.
Enumeration of fecal indicator bacteria
E. coli and Enterococcus spp. numbers in tank water samples were enumerated using membrane filtration method (US EPA, 1997; US EPA, 2002). Sample serial dilutions were made and filtered through 0.45 mm pore sized (47 mm diameter) nitrocellulose membranes (Millipore, Tokyo, Japan). The membranes were placed on modified mTEC agar (Difco, Detroit, MI) and membrane-Enterococcus indoxyl-b-D-glucoside (mEI) agar (Difco) for the isolation of E. coli and Enterococcus spp., respectively. Modified mTEC agar plates were incubated at 35 C for 2 h to recover stressed cells, followed by incubation at 44 C for 22 h (US EPA, 2002), and mEI agar plates were incubated at 41 C for 48 h (US EPA, 1997).
2.3.
Concentration of tank water samples
Approximately, 19 L water sample from each rainwater tank was concentrated by hollow-fiber ultrafiltration system (HFUS), using Hemoflow HF80S dialysis filters (Fresenius Medical Care, Lexington, MA, USA) as previously described (Hill et al., 2005). Each sample was concentrated to approximately 100 mL. Each 100 mL sample was further centrifuged at 3000 g for 30 min at 4 C. The supernatant was discarded, and the pellet was resuspended in 5 mL of sterile distilled water.
363
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Table 1 e - Primer sequences for quantitative PCR (qPCR) assays for the quantitative detection of opportunistic pathogens. Pathogens
Primer sequence (50 e30 )a
A. hydrophila
F e AAC CTG GTT CCG CTC AAG CCG TTG R e TTG CTC GCC TCG GCC CAG CAG CT F e GCA ATG TCA ACA GCAA R e CAT AGC GTC TTG CATG F e TGC TGG TGG CAC AGG ACA T R e TTG TTG GTG CAG TTC CTC ATT G F e CGT ATT AGC AGA GAG CCA ACC A R e GTG ATT TTA CTC GCT TTG TGC AA
L. pneumophila P. aeruginosa S. aureus a
Genes
Functions
References
lip
Novel lipase
Casco´n et al., 1996
mip
Surface protein virulence factor
Wilson et al., 2003
regA
Toxin A synthesis regulating gene
Shannon et al., 2007
sec
Enterotoxin C gene
Shannon et al., 2007
F, forward primer; R, reverse primer.
2.4.
DNA extraction
For quantitative PCR (qPCR) analysis, DNA was extracted from the pellet obtained from 1.5 mL (equivalent to 5.7 L of water) of concentrated samples using a DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA). All DNA samples were quantified using a NanoDrop spectrophotometer (ND-1000, NanoDrop Technology, Wilmington, DE). Each DNA sample was amplified using a universal bacterial PCR assay to confirm successful DNA extraction process (Boon et al., 2000).
2.5.
Preparation of qPCR standard curves
To obtain qPCR standard curves, DNA was extracted from the 1 mL of broth cultures of A. hydrophila ATCC 7966, L. pneumophila ATCC 33152, P. aeruginosa ATCC 15629, and S. aureus ATCC 12600 using DNeasy Blood and Tissue Kit (Qiagen). The concentrations of genomic DNA samples were determined using a NanoDrop spectrophotometer (ND-1000, NanoDrop Technology). The gene copies were calculated, and a 10-fold dilution ranging from 106 to 101 copies per mL of DNA extract was prepared from the genomic DNA, and stored at 20 C until use. For each standard, the genomic copies were plotted against the cycle number at which the fluorescence signal increased above the threshold value (CT value).
2.6.
qPCR primers and conditions
The primers used for qPCR assays are shown in Table 1. qPCR amplifications were performed in 20-mL reaction mixtures using Sso Fast EvaGreen Supermix (Bio-Rad Laboratories, Calif). The PCR mixture contained 10 mL of Supermix, 300e500 nM each primer, 5 mL of template DNA and DNaseand RNase-free deionized water. The qPCR conditions were as follows: 15 min at 95 C, 45 cycles of 60 s at 94 C, 60 s at 62 C, and 90 s at 72 C (for A. hydrophila), 10 min at 95 C, 45 cycles of 15 s at 95 C, 60 s at 60 C (for P. aeurginosa and S. aureus), and 15 min at 95 C, 45 cycles of 30 s at 94 C, 60 s at 54 C, and 60 s at 72 C (for L. pneumophila). For each qPCR experiment, a negative control (sterile water) was included. The qPCR assays were performed using the Bio-Rad iQ5 (Bio-Rad Laboratories). To separate the specific product from non-specific products, DNA melting curve analysis was performed for each qPCR assay.
2.7. qPCR assays reproducibility and lower limit of detection (LLOD) The reproducibility of the qPCR assays was assessed by determining intra-assay repeatability and inter-assay reproducibility. The coefficient of variation (CV) was calculated using six dilutions (106 to 101 gene copies) of genomic DNA isolated from opportunistic pathogens. Each dilution was quantified in replicates. The CV for evaluation of intra-assay repeatability was calculated based on the CT value by testing the six dilutions six times in the same experiment. The CV for inter-assay reproducibility was calculated based on the CT values of six dilutions on three different days. To determine qPCR LLOD, extracted genomic DNA samples of opportunistic pathogens were quantified using a NanoDrop spectrophotometer (ND-1000, NanoDrop Technology). Ten-fold serial dilutions were made and tested with the qPCR assays. The lowest amount of DNA detected in replicate assays was considered qPCR LLOD.
2.8.
Recovery efficiency
To determine the method recovery efficiency, E. coli was used as a representative microorganism. The recovery efficiency of E. coli was assumed to be similar to those opportunistic pathogens tested in this study. Nineteen liters deionized water (n ¼ 3) and 19 L tank water (n ¼ 3) samples were spiked with known cells of E. coli. In brief, E. coli (ACM 1803) was grown overnight in Nutrient Broth (Oxoid, UK), and the number of cells in the broth culture was determined using a spread plate method. Ten-fold serial dilutions (2.1 106, 2.1 105, 2.1 104) were made and seeded into the distilled water and tank water samples. The samples were concentrated, and DNA extraction was performed according to the method described above. All DNA samples were tested in triplicate with qPCR of E. coli 23S rRNA gene (Chern et al., 2011). The recovery efficiency (%) was calculated using the following equation: recovery (%) ¼ [Genomic units (GU) of E. coli 23S rRNA after qPCR analysis/seeded GU of E. coli 23S rRNA before concentration process] 100.
2.9.
PCR inhibitors
An experiment was conducted to determine the potential presence of PCR inhibitors in tank water samples collected randomly from 10 different tanks. Ten-fold serial dilutions
364
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Table 2 e The amplification efficiency, correlation coefficient (r2), intra-assay and inter-assay coefficient of variation (CV) for the qPCR assays within the range of 106 to 102 GU of opportunistic pathogens. Opportunistic pathogens
Amplification efficiency (%)
r2
98 106 98 103
0.99 0.98 0.99 0.95
Mean SD of CV Intra-assay
A. hydrophila L. pneumophila P. aeruginosa S. aureus
0.86 0.84 0.94 0.82
0.1 0.1 0.1 0.2
Inter-assay 3.1 4.3 3.6 2.9
0.7 0.6 1.0 0.9
SD: Standard deviation.
were made for each sample. All DNA samples (undiluted, 10fold diluted, and 100-fold diluted) were seeded with 103 gene copies of human specific HF183 Bacteroides markers (Seurinck et al., 2005). The CT values obtained for the HF183 spiked DNA samples were compared to those of the distilled water to obtain information on the level of PCR inhibitors.
2.10.
Statistical analysis
occurring at lower cell counts (seeded). The mean recovery efficiency was 72% 9%. The estimated recovery efficiency in tank water samples ranged between 55% and 71% with the greatest variability occurring at lower cell counts (seeded). The mean recovery efficiency was 60% 8%.
3.3.
PCR inhibitors
Fecal indicator bacteria and opportunistic pathogen data were log10 transformed. The nonparametric Spearman rank correlation was performed to obtain correlation between CFU of E. coli and Enterococcus spp. A binary logistic regression analysis was also performed to obtain correlations between the presence/absence results of opportunistic pathogens and the CFU of E. coli and Enterococcus spp. A difference was considered significant if the P value for the model chi-square was 0.05 and the confidence interval for the odds ratio did not overlap. IBM SPSS software (version 21.0) was used for all statistical analysis.
The mean CT value for the HF183 spiked distilled water DNA was 23.8 0.4. For tank water samples, the mean CT value was 22.8 0.5 when undiluted DNA was spiked with the HF183. For 10-fold and 100-fold dilutions of DNA spiked with the HF183, these values were 23.2 0.6 and 23.8 0.3, respectively. Analysis of variance (ANOVA) was performed to determine the differences between the CT values obtained for distilled water and those obtained for tank water samples. No significant differences were observed between the CT values for the HF183 spiked distilled water, undiluted DNA, and serially diluted DNA (10-fold and 100-fold), thus indicating that the tested tank water samples were free of PCR inhibitors.
3.
Results
3.4.
3.1.
qPCR standards, assay reproducibility and LLOD
Culturable E. coli exceeded that specified by Australian Drinking Water Guidelines for 74% of 72 tank water samples. The Colony forming units (CFU) of E. coli in positive samples ranged between 0.5 and 3.7 log10 CFU per 100 mL of water with an average of 1.8 log10 CFUs per 100 mL (Fig. 1). Similarly, 94% samples were also positive for Enterococcus spp., and the
qPCR standards for target opportunistic pathogens were analyzed in order to determine the reaction efficiencies. The standard curves had a linear range of quantification from 106 to 101 gene copies per mL of DNA extracts. The amplification efficiencies ranged from 98% to 106%, and the correlation coefficient (r2) ranged from 0.95 to 0.99 for all qPCR assays. The mean intra-assay and inter-assay coefficient of variation (CV) values and standard deviations (SD) were less than 1% and 5%, respectively, indicating high reproducibility (Table 2). LLOD assays were performed using purified genomic DNA isolated from pure culture of each opportunistic pathogen. To determine the reproducibility of the assays, three replicates were tested for each target pathogen. The real-time PCR LLOD ranged from 2 to 8 fg for L. pneumophila, P. aeruginosa and S. aureus. For A. hydrophila, the LLOD was 13 fg.
3.2.
E. coli and Enterococcus spp. numbers
Recovery efficiency
The recovery efficiency was determined by seeding tank water and distilled water samples with known numbers of E. coli. The estimated recovery efficiency in distilled water samples ranged between 65% and 85% with the greatest variability
Fig. 1 e Box-and-whisker plots of the log10 CFU of fecal indicator bacteria in tank water samples. The inner box lines represent the medians, while the outer box lines represent the 25th and 75th percentiles.
365
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Table 3 e Percentage of tank water samples that were positive and quantifiable for opportunistic pathogens tested in this study. Opportunistic pathogens A. hydrophila L. pneumophila P. aeruginosa S. aureus a b
No. of positive samples/no. of total samples tested (% of positive samples)a
No. of quantifiable samples/no. of total samples tested (% of quantifiable samples)b
21/72 (29) 5/72 (6) 9/72 (13) 11/72 (15)
14/72 (19) 5/72 (6) 4/72 (6) 10/72 (14)
: Real-time PCR. : qPCR.
numbers ranged between 0.3 and 3.6 log10 CFUs with an average of 1.5 log10 CFU per 100 mL. Among the 72 tank water samples tested, 72% were positive for both E. coli and Enterococcus spp., and 97% were positive for either E. coli or Enterococcus spp. numbers. A nonparametric Spearman rank correlation was performed to determine whether any correlation exists between E. coli and Enterococcus spp. in tank water samples. Significant correlation was found between these two fecal indicator bacteria (Spearman’s rs ¼ 0.50; P < 0.001).
S. aureus (2.8e4.6 log10 GU with an average 3.8 log10 GU) per 100 mL of water (Fig. 2). Binary logistic regression analysis was used to identify whether any correlation existed between the numbers of E. coli or Enterococcus spp. and the presence/ absence results for the opportunistic pathogens (Table 4). The presence/absence results of the opportunistic pathogens did not correlate with E. coli and Enterococcus spp. numbers.
4. 3.5. Genomic units (GU) of opportunistic pathogens in tank water samples Of the 72 tank water samples tested, A. hydrophila was the most prevalent (29%) among all the opportunistic pathogens tested in this study (Table 3). The prevalence of other opportunistic pathogens was as follows: 15% samples were positive for S. aureus, 13% were positive for P. aeruginosa. The prevalence of L. pneumophila (6%) was low compared to other opportunistic pathogens. Of the 72 tank samples, 8% were positive for two opportunistic pathogens, and 32% were positive for at least one opportunistic pathogen. In contrast, none of these opportunistic pathogens was detected in 56% tank water samples. A number of samples although gave PCR positive signal but was not quantifiable hence, not included in the qPCR analysis. The range of GU was as follows: A. hydrophila (2.1 and 4.5 log10 with an average 3.5 log10 GU), L. pneumophila (3.2e4 log10 GU with an average 3.4 log10 GU), P. aeurginosa (1.5e4.3 log10 GU with an average 3.4 log10 GU) and
Discussion
In this study, roof-captured rainwater samples were tested for the numbers of E. coli and Enterococcus spp., using conventional culture-based methods. In addition, qPCR assays were used to obtain information on the GU of four opportunistic bacterial pathogens. The application of PCR-based methods has generated interest in the direct monitoring of pathogens in water because they are rapid, sensitive and can detect microorganisms that are difficult to grow using conventional culture methods. In view of this, qPCR assays were used to quantify opportunistic pathogens in tank water samples. The detection sensitivity of the qPCR assays used in this study was evaluated by amplifying known concentrations of DNA from target opportunistic pathogens. The LLOD ranged from 0.2 to 13 fg for the tested opportunistic pathogens indicating that the detection sensitivity values of our qPCR assays were comparable to the values reported in the literature (Behets et al., 2007; Sails et al., 2002). Prior to setting up qPCR assays, specificity of each primer was determined by searching for similar sequences in microbial genomes using the Basic Local
Table 4 e Correlations between colony forming units (CFUs) of fecal indicator bacteria and the presence/ absence of opportunistic pathogens.a Relationship E. coli vs. A. hydrophila E. coli vs. L. pneumophila E. coli vs. P. aeurginosa E. coli vs. S. aureus Enterococcus spp. vs. A. hydrophila Enterococcus spp. vs. L. pneumophila Enterococcus spp. vs. P. aeurginosa Enterococcus spp. vs. S. aureus
Fig. 2 e - Box-and-whisker plots of the log10 GU of opportunistic pathogens in tank water samples. The inner box lines represent the medians, while the outer box lines represent the 25th and 75th percentiles.
a
R2
P value
Odd ratio
0.027 0.001 0.011 0.006 0.004 0.056 0.071 0.001
0.275 0.854 0.481 0.568 0.660 0.454 0.095 0.845
1.000 1.000 1.000 1.000 1.000 0.994 1.001 1.000
Data are results of binary logistic regression analysis (P values) performed to identify correlations between the CFU of fecal indicator bacteria and the presence/absence of opportunistic pathogens in roof-captured rainwater samples.
366
w a t e r r e s e a r c h 5 3 ( 2 0 1 4 ) 3 6 1 e3 6 9
Alignment Search Tool (BLAST) program (http://www.ncbi. nlm.nih.gov/BLAST/). The cross-reactivity of each primer set for each target has also been assessed against a panel of 20 microorganisms (data not shown). The primers did not amplify any PCR products other than those products that were expected, suggesting a high specificity for the primers used in this study. In the 72 tank water samples tested, 74% samples had >1 CFU E. coli per 100 mL of water, exceeding Australian Drinking Water Guidelines (ADWG, 2011). The numbers of E. coli and Enterococcus spp. were highly variable in the tank water samples, around 19% and 26% tank water samples had >2 log10 CFU E. coli and Enterococcus spp. per 100 mL of water, suggesting the occurrence of fecal pollution from birds, mammals insects and perhaps environmental sources (Ahmed et al., 2012a, 2012b). This was not unexpected because the samples were collected within one to four days after rainfall events, when fecal and other organic matter deposited on the roof enters tanks via roof runoff. Enterococcus spp. was more prevalent than E. coli in tank water samples. A number of samples (17 of 72) were culture positive for Enterococcus spp. but negative for E. coli, possibly because Enterococcus spp. persist longer in the tank water than E. coli (Ahmed et al., 2013). It is also possible that all Enterococcus spp. isolated from tank water samples were not from a fecal origin. A recent study reported the presence of soil or plant related Enterococcus mundtii and Enterococcus casseliflavus in a number of tank water samples in Southeast Queensland, Australia (Ahmed et al., 2012b). In this study, opportunistic pathogens were detected in one or more tank water samples. The overall prevalence of S. aureus, P. aeruginosa, and A. hydrophila (ranged from 13 to 29%) in tank water samples was higher than L. pneumophila (6%). The occurrence of L. pneumophila in this study, however, was lower than previously reported studies (Albrechtsen, 2002; Broadhead et al., 1998). The occurrence of A. hydrophila in this study is comparable to previous studies undertaken in Australia and New Zealand (Simmons et al., 2001; CRC for Water Quality Treatment, 2006). The occurrence of pathogens in tank water samples in different studies is expected to vary due to several factors such as sample volume, detection method sensitivity, climatic conditions and physical parameters of the rainwater system. Another important factor which may also influence the occurrence of pathogens is the types of wildlife that has access to roof. In Southeast Queensland, for example, marsupial possums are a common animal in and around urban dwellings and frequently traverse roof tops in their nocturnal movements. Birds and wildlife including possums and other animals, therefore, could be a major source of certain bacterial pathogens in tank water in Southeast Queensland (Ahmed et al., 2012b). In this study, one or more opportunistic pathogens were quantitatively detected in a number of nonpotable tank water samples. It can be postulated that the magnitude of health risks from nonpotable uses could be lower than potable use due to lower exposure levels (Ahmed et al., 2010). While much of the focus on bacterial pathogens in drinking water is predominantly on enteric indicators (E. coli) and pathogens (Salmonella spp.) in tank water, opportunistic pathogens tested in
this study are also considered to be of concern for human health. In this study, the high level of GU (average 3.2 to 3.8 log10 per 100 mL) of opportunistic pathogens along with E. coli and Enterococcus spp. in potable tank water samples indicate a poor level of microbial quality, and could represent potential health risks to users with low immunity. The result is particularly of significance because a recent study has demonstrated the inefficacy of the under sink filtration methods that are being used for tank water purification prior to drinking in Southeast Queensland, Australia (Ahmed et al., 2012a). There have been concerns that the presence of these opportunistic pathogens in drinking water may represent public health risks (Egorov et al., 2011; Farkas et al., 2012). High numbers of S. aureus (average 3.8 log10) were detected in certain tank water samples. S. aureus is known to cause food poisoning, a variety of skin abscesses, pustules and septicemia (LeChevallier and Seidler, 1980). A. hydrophila is considered to be of concern because of the high GU (average 3.5 log10) detected per 100 mL of water in this study. A recent study demonstrated a link between Aeromonas spp. isolated from clinical and water samples, indicating a transmission from water (Khajanchi et al., 2010). The high GU of L. pneumophila, in certain tank water samples can be considered to be of concern because direct exposure to aerosols containing Legionella spp. generated from cooling towers, spa and shower heads are implicated as the vehicle for bacterial transmission (Armstrong and Haas, 2007a; Bauer et al., 2008; Schoen and Ashbolt, 2011). We acknowledge that qPCR results do not provide information regarding the infectivity status of the target microorganisms. Another major limitation of PCR is its inability to distinguish between viable and nonviable pathogens. Therefore, over estimation of opportunistic pathogens cannot be ruled out (Krøjgaard, 2011). In contrast, several method comparison studies also reported high numbers of qPCR quantified E. coli and Enterococcus spp. cell equivalents compared to culture-based CFU in recreational waters (Converse et al., 2011; Haugland et al., 2005). A possible explanation for such discrepancy is that certain microorganism including pathogens can enter viable but not culturable (VBNC) state when subjected to adverse environmental conditions (Kolling and Matthews, 2001; Heim et al., 2002). In addition, culture-based methods may underestimate the bacterial numbers due to presence of injured or stressed cells (Juhna et al., 2007). Another important factor that needs to be recognized is that some target opportunistic pathogens tested in this study have high ID50 levels. For example, large oral doses up to 1010 CFU, A. hydrophila failed to produce diarrhea in human volunteers (Morgan et al., 1985). Therefore, the level of A. hydrophila GU that were detected in tank water samples may not pose any risks to healthy humans, however, immunocompromised and elderly population could be at risk. For L. pneumophila, a small dose such as
