Adf water final report
Maintaining Food Safety by Protecting Irrigation Water from Faecal Contamination
Saskatchewan Ministry of Agriculture ADF Project
Final Report, February 1, 2010
Prepared by Chris Yost
Chris YostCanada Research ChairBiology Department, University of Regina3737 Wascana ParkwayRegina, SK S4S0A2chris.firstname.lastname@example.org
Water quality in Saskatchewan is directly impacted by a variety of human activities.
Agricultural and municipal activities can adversely affect water quality if failures in waste management cause faecal material to be transported into aquatic ecosystems. The resultant pollution can cause
illnesses and have lethal consequences when enteric pathogens are present in the faecal material. Therefore, our ability to identify and eliminate faecal contamination of water, now and in the future, is essential to reduce the risk of waterborne disease. Microbial source tracking (MST) is a promising
recently emerged technology that can identify animal specific sources of faecal pollution. The potential of MST technology is appealing as it could facilitate rational mitigation strategies to prevent future
microbial contamination events, particularly in watersheds where faecal contamination may be the result of both municipal and agricultural inputs. This study first developed and validated a MST strategy in the laboratory and subsequently used the Qu'Appelle Valley, between Regina and Craven as
a study site to further validate the approach in a Saskatchewan context. The section of the Qu'Appelle River under study is an important source of irrigation water for several vegetable producers who
provide fresh produce to southern Saskatchewan consumers, and protection of the irrigation water quality is of critical importance for food safety and maintaing the excellent reputation of Saskatchewan producers as suppliers of high quality produce. Furthermore, the study area was selected due to the
influence of both intensive agricultural activities and municipal development on water quality in the region.
A Bacteroidales-based microbial source tracking strategy was selected for development. The
results of the project clearly illustrate the success of this method in accurately determining host specific
sources of faecal contamination. In particular, markers for the specific detection of human, ruminant, pig, and bovine, faeces have been validated and new markers for horse and Canada Goose faeces have
been developed for use in Saskatchewan. These markers provide both sensitive and specific detection of host specific faecal material. The testing of the MST markers in the Qu'Appelle River and Wascana Creek has clearly demonstrated the benefit of using a combined measurement of both alternative and
conventional faecal indicators (Eschericheria coli
) for a better understanding of the sources and severity of faecal contamination events. In addition to the Bacteroidales-based MST approach, a
protocol was developed for the detection of antibiotics in water sources influenced by agricultural and urban inputs. The results of this study suggest that detection of particular antibiotics, such as erythromycin may provide additional useful information when measuring water quality and inputs from
wastewater systems. In conclusion, the results of this study support the proposal that bacterial source tracking can provide Provincial government agencies and researchers with effective new tools for
comprehensive studies to characterize faecal pollution inputs in Saskatchewan watersheds. Ultimately adoption of these tools can help to increase the effectiveness of identifying and remediating faecal pollution in Saskatchewan water sources and ultimately help protect irrigation water quality in the
Irrigation, especially in dry climate agriculture, is critical to successful crop production.
Frequently, irrigation water is derived from open sources, such as rivers that have the potential to be
impacted by faecal pollution. Pollution of irrigation water by faecal material can have lethal
consequences, when pathogens, such as enterohemorragic Eshericheria coli
spp. are transmitted into the water along with the deposited faecal material
(Sivapalasingam et al., 2004; Steele and Odumeru, 2004; CDC, 2008). Contaminated irrigation water
has been suspected in various cases of human infections associated with consumption of fresh or
minimally processed fruits and vegetables (Steele and Odumeru 2004). A recent significant example in
North America was the 2008 Salmonella
serotype Saintpaul outbreak in the USA that was traced to
peppers irrigated with Salmonella
contaminated water. The consequence was one of the largest
nationwide outbreaks of food-borne illness in the USA (CDC, 2008). These large outbreaks emphasize
the importance of efficiently protecting open source irrigation water from faecal pollution and
subsequently monitoring the quality of irrigation water routinely. The guidelines regarding microbial
water quality for irrigation water vary depending on the regulatory agency. A limit of 1000 faecal
coliforms per 100 ml of water has been recommended by the World Health Organization (WHO) for
unrestricted irrigation (Shuval 2007), the Canadian Water Quality Guidelines for Irrigation (CWQGI)
has recommended more stringent guidelines and suggests a maximum allowable count of 100 faecal
coliforms per 100 ml and 1000 total coliforms per 100 ml of irrigation water (Jones and Shortt 2005).
The use of faecal coliform monitoring to assess irrigation water quality will remain an important tool
in irrigation water quality assurance. However, this tool has significant limitations in helping vegetable
producers reduce or eliminate faecal contamination of open sources of irrigation water. Sources of
faecal contamination are not indicated when using conventional E. coli
detection procedures, and
source identification can be of great value to a vegetable producer in remediating the contamination.
Furthermore, the consistent reports of long term persistence of E. coli
within aquatic systems may limit
the ability of E. coli
detection in identifying recent faecal contamination events (Van Elsas et al., 2011;
Ishii and Sadowsky, 2008). Microbial source tracking (MST) is a promising recently emerged
technology (Domingo et al., 2007; Field and Samadpour 2007) that may assist in developing and
validating rational mitigation strategies to prevent future microbial contamination of irrigation water
sources. Despite the recent growth in MST research, development and validation of these new faecal
indicators are required before MST can be used with confidence to identify and reduce faecal
contamination of irrigation water sources. This project sought to determine the potential of a
Bacteroidales MST approach for measuring faecal pollution in the Qu'Appelle River, in an area used
for irrigation of fresh leafy green produce. Bacteroidales are faecal anaerobes that provide a suitable
alternative to coliform bacteria (e.g. E. coli
) as indicators of faecal contamination in water for a variety
of reasons. Bacteroidales are present in faecal material at concentrations 1000-fold greater than
coliform species; making them a potentially highly sensitive indicator of faecal contamination (Fiksdal
et al. 1985). Additionally, because Bacteroidales are anaerobic, they have a poor survival rate outside of
the intestinal environment (Fiksdal et al. 1985) and this is useful in predicting future contamination
events since their viable presence in the environment is an unambiguous indicator of ongoing
contamination. Bacteroidales MST has only recently been developed and remains to be tested in
Saskatchewan for irrigation water protection. Because the Bacteroidales
approach is relatively new its
efficacy has been investigated at limited locations and little is known regarding the geographical
robustness of host specificity within the Bacteroidales. As well, the persistence and transport of
in the environment needs to be vigorously quantified to estimate the spatial and temporal
range over which faecal inputs can be detected. Detection of the Bacteroidales host-specific gene
markers can be accomplished via conventional end-point polymerase chain reaction (PCR) which
provides a qualitative presence or absence result. Quantitative PCR (qPCR) approaches have also been
developed which provides a quantitative measure of the abundance of the Bacteroidales gene marker.
Both approaches were tested in this study.
Tracking sources of faecal contamination can also involve non-biological approaches, such as
determining the presence of particular chemicals linked to specific sources of contamination (Blanch et
al., 2006). This study examined the detection of veterinary antimicrobials as a possible indicator for
water quality. Earlier work of co-applicant, A. Cessna, has involved the detection of agriculturally
related chemicals within watersheds (Kuchta and Cessna, 2009; Cessna et al., 2006). Detection of
veterniary antimicrobials may provide an additional tool to determining sources of faecal
contamination in agricultural watersheds. Swine, poultry and cattle are increasingly being produced in
large confined animal feeding operations. With such production, antimicrobials are administered to
prevent disease and to promote growth by increasing the rate of weight gain. Since many antimicrobials
are poorly absorbed through the gut, they can be excreted up to 80% or more in the faeces and urine.
Recently, other researchers have shown that antimicrobials used in the livestock industry can be
detected in surface waters in Canada (Forrest et al., 2006), the United States (Campagnolo et al., 2002)
and Europe ( Hirsch et al., 1999). Since faecal contamination frequently enters the Qu' Appelle
watershed (Fremaux et al 2009a) it is important to determine if antimicrobials are also entering the
watershed. Detection of veterinary antimicrobials would imply that animal waste may be a factor in
the faecal contamination and the information, combined with data from the microbial source tracking,
can be used to identify the potential sources of contamination.
The study site for the project includes a portion of the Qu'Appelle Valley watershed (52 000km2).
The Qu'Appelle River extends over 400 km from headwaters near Lake Diefenbaker to join the
Assiniboine River in Manitoba. More than 95% of the drainage area is composed of agricultural fields
(wheat, barley, canola) and cattle pastures, but the Qu'Appelle River also runs through urbanized areas
such as the towns of Lumsden and Craven (Hall et al. 1999). A series of seven dams and control
structures along the river have created reservoirs such as Buffalo Pound Lake, Pasqua Lake and
Katepwa Lake along the Qu'Appelle River valley. The discharge regime of the Qu'Appelle River is
thus a mix of the regimes of a typical prairie river and lake drainage, and is modified by water additions
from the south Saskatchewan River system at Lake Diefenbaker (Pomeroy et al. 2005). Within the
selected study site the Qu'Appelle River is under the influence of Wascana Creek, which receives
effluent from the Regina city municipal wastewater treatment plant. The study area also includes
locations where water is drawn from the Qu'Appelle River for irrigation of fresh leafy vegetables.
Contributing sources of contamination to the Qu'Appelle River include diffuse inputs from wildlife,
source specific inputs from livestock operations, rainfall runoff events, as well as effluent discharges of
intermittent poor quality and unidentified storm water and⁄or sanitary sewer discharges of unpredictable
nature (Water Pollution Control Branch 1984). The detection of E. coli
levels above the recommended
limit of 100 CFU per 100 mL (Jones and Shortt, 2005) within the study area range from 0% to 100%
since sampling began in 2005. The frequent and variable inputs of faecal material into the sampling
sites (Fremaux et al., 2009a) confirmed its suitability for field testing of the Bacteroides
approach. Combined with the long history of consistent land use and descriptions of the hydrology it is
an excellent site to study the efficacy of newly developed MST tools.
2.1 Study Site and water sample collection:
Sites along the Wascanna Creek and Qu'Appelle river from Regina to Craven were sampled at
weekly intervals during the growing seasons of 2008, and 2009. The original proposal was to sample in
2009 and 2010, however due to efficient results obtained in the lab we were able to begin our sampling
earlier than planned. Figure 1 provides a schematic of the sampling area. All samples were stored at 4 C
and were delivered to the lab for subsequent analysis within 3 hours of collection. Weekly sampling
was not conducted in 2010 because the successful large sampling efforts of 2008 and 2009 resulted in
sufficiently robust data to evaluate the Bacteroidales
based markers. Therefore in 2010 we were able to
improve upon the original proposal by adding a study on persistence of the Bacteroidales faecal
indicator markers (see methods section 2.5 for more detail).
Water samples collected in the field, were processed in the lab for subsequent PCR analysis as
follows: Samples were concentrated by filtering 400 ml of water through a 0.45 µm pore size
membrane using a filter funnel and vacuum system (Millipore Inc., USA). For each sample, the filter
was placed in a sterile 15 ml falcon tube (Fisher Scientific, Ottawa, ON, Canada) containing 5 ml of
sterile distilled water. The tubes were thoroughly vortexed and 4 ml of this solution was centrifuged at
11,000 g for 3 min. The pellet was resuspended in 200 µl of sterile distilled water and DNA was
extracted from this suspension by using the Power Soil DNA kit (MoBio Laboratories Inc.), according
to the manufacturer's instructions.
2.2 Faecal material collection for testing sensitivity and specificity of host specific Bacteroides
To determine the specificity and sensitivity of the different Bacteroides
primer sets (Table 1)
designed for detection using conventional PCR, a total of 265 individual fresh faecal samples were
aseptically collected from 12 different host groups and 8 human derived raw sewage samples were
obtained from 4 localities in Saskatchewan (Prince Albert, Lloydminster, North Battleford and Regina).
Individual human specimens (n
= 54) were provided by the Disease Control Laboratory of
Saskatchewan. Cow (n
= 51), chicken (n
= 21) and part of the pig faeces (n
= 11) were collected as
certainly as possible from separate animals from southern SK farms. Additional faecal samples from
= 39) were obtained from the Prairie Swine Centre (Saskatoon, SK). Faecal samples from other
animal species (n
= 80) were taken from the Calgary zoo (Alberta), except the goose samples (n
that were collected from the Wascana Lake (Regina, SK) (Table 2). Samples were transported to the
laboratory on ice and kept at 4°C until the time of analysis. Additional samples from similar locations
were obtained for testing of the quantitative PCR Bacteroidales markers (Table 5)
DNA was directly extracted from 0.25 g of faecal material or 0.25 ml of raw sewage by using
the Power Soil DNA kit (MoBio Laboratories Inc., Carlsbad, CA, USA) according to the
manufacturer's instructions. Water samples from the Qu'Appelle River were first concentrated by
filtering 100 ml of water through a 0.45 !m pore size membrane using a filter funnel and vacuum
system (Millipore Inc., USA). For each sample, the filter was placed in a sterile 15 ml falcon tube
(Fisher Scientific, Ottawa, ON, Canada) containing 5 ml of sterile distilled water. The tubes were
thoroughly vortexed and 2 ml of this solution was centrifuged at 11,000 g for 3 min. The pellet was
resuspended in 200 !l of sterile distilled water and DNA was extracted from this suspension by using
the Power Soil DNA kit (MoBio Laboratories Inc.), according to the manufacturer's instructions. The
DNA was used in the PCR assays as described below in section 2.4.
2.3 Enumeration of E. coli
by using the Colilert 18/Quanti-Tray system.
The Colilert method was used to process river water samples. Each undiluted or appropriate
diluted water sample (100 ml) was added to the Colilert 18 reagent and poured into a 97-well Quanti-
Tray system, sealed and incubated at 37°C for 22h according to the manufacturer's instructions
(IDEXX). Total coliforms and E. coli
were simultaneously detected and enumerated by counting the
number of yellow wells and yellow/fluorescent wells under UV light, respectively. The most probable
number (MPN) could be estimated from a chart provided by the manufacturer. Because E. coli
more specific indicator of faecal contamination (Buckalew
et al., 2006), only data observed for these
bacteria are considered (results were reported as MPN per 100 mL).
2.4 Conventional PCR and Quantitative PCR procedure for detection of host-specific Bacteroidales
bacterial pathogens in the Qu'Appelle Valley
Conventional End Point PCR assays with primers designed for total, human, ruminant and pig-
markers was applied on faecal (including raw sewage) and water samples using
primer pairs previously designed by Bernard and Field (2000a, 2000b) (Table 1). All amplification
reactions were processed in the MyCycler Thermal Cycler (Bio-Rad Laboratories, Hercules, CA,
USA), and carried out in a total volume of 25 !l reaction mixture containing 5 µl of template DNA, 2
mmol l-1 MgSO4, 0.2 mmol l-1 of each deoxynucleoside triphosphate (Invitrogen, Carlsbad, USA), 0.4
µmol l-1 of each primer (Sigma, Ontario, Canada), 1X Taq reaction buffer, 3 !g of Bovine Serum
Albumine (Sigma, Ontario, Canada) and 1 U of Taq DNA polymerase (UBI, Calgary, Canada). The
amplification program used was as follows: 5 min initial denaturation of DNA at 95°C, followed by 30
cycles of 30s denaturation at 94°C, 30s primer annealing at the temperature specific for each primer
pair (Table 1), and 1 min extension at 72°C. Amplification was completed by a final extension step at
72°C for 6 min. PCR products were separated by gel electrophoresis on 1.5% agarose and visualized by
ethidium bromide staining.
Real-time Quantitative PCR (qPCR) detection of total, human, ruminant and bovine-specific
markers was applied on water samples using primer and probe sets previously designed
(Table 2). To estimate horse faecal pollution in the water samples, a new primer and probe set was
developed from a previously identified horse-specific 16S rRNA Bacteroidales
sequence. All real-time
PCR assays were conducted on an iQ5 Real-Time PCR Detection system (Bio-Rad Laboratories).
Annealing temperatures as well as primer and probe concentrations were experimentally optimized to
obtain optimal performance for specificity and sensitivity. The reaction mixture contained 4 µl of
template DNA, appropriate concentrations of primers and probe (Table 4), 5 µg bovine serum albumin
(Promega, WI, USA), 12.5 µl iQ Supermix (Bio-Rad Laboratories) and 25 µl water q.s.p. The
amplification program used for the Hum-Bac
assays was as follows: 95°C for 6 min,
followed by 40 cycles of 95°C for 15 s, 61°C (Hum-Bac
) or 60°C (Rum-Bac
) for 15 s and 72°C for 45
s. Amplification program for the All-Bac
assays consists of: 95°C for 6 min,
followed by 40 cycles of 95°C for 15 s (Bov-Bac
), 20 s (Hor-Bac
) or 30 s (All-Bac
) and 60°C for 45 s
) or 1 min (Bov-Bac
). Standard curves for all different qPCR assays were
generated using serial dilutions of recombinant plasmid DNA containing known copy numbers of
appropriate target genes. The qPCR negative controls (no DNA) as well as plasmid DNA standards
were always performed in triplicate and DNA template from water samples were processed in
duplicate. Furthermore, a synthetic DNA internal amplification control (IAC) was included in all Bov-
qPCR reactions in order to confirm that false-negative results due to Taq inhibitors that could be
present in the DNA extract did not occur. The IAC used in this study, with the sequence
AAGGAACGCAACAAGC-3', was slightly different than that described by Shanks et al. (2008). It
contained the same primer binding sequences as the bovine-specific Bacteroidales
DNA target, plus a
sequence that hybridized with a probe labelled with the TET fluorophore in order to distinguish
amplification of the IAC DNA sequence from that of the target. An IAC concentration of 10-10 µM was
determined as the most appropriate since it does not compete with the target amplification.
2.5 Persistence of the Bacteroides markers
in the environment
The persistence study was conducted in the Wascana Creek, upstream of the city of Regina.
Three trials were preformed during the summer of 2010, with each trail lasting between fourteen and
seventeen days. Two pieces of rebar were placed one meter apart within the flow of the creek. Bottles
were secured on wire with plastic ties and positioned half a meter below the surface of the water.
HOBO Pendent Data Logger's were attached to the rebar to collect temperature and light data (Onset
Computer Corporation, Bourne, MA). In trials one and two duplicate replicates were placed in the
Individual faecal samples, cows, were collected from a local location. Human waste water,
primary influent, was collected from the Regina Waste Water Treatment Plant. DNA was extracted
from all cow samples to confirm the presence of genetic markers then frozen at -20C. Primary influent
was collected three days prior to the start of each trial and stored at -4C. Five grams of individual faecal
sample was weighed out and added to 700 ml of sterile water. The contents were homogenized and
debris was allowed to settle for 10 minutes. 100ml of supernatant was added to 500ml sterile plastic
bottles. Waste water was mixed and 40ml was added to each bottle. In the field 160ml of river water
was aseptically added to each bottle, bringing the final volume to 300ml. Water samples were collected
every day for the first 5 days and then twice per week until the host specific Bacteroidales
were below the limit of detection using qPCR. DNA was extracted from the water samples as described
in section xx and qPCR reactions were as per described in section 2.1. E. coli enumeration was
performed using the Colilert-18/Quanti-Tray system as described in methods section 2.3.
2.6 Detection of agriculturally related antimicrobials in the Qu'Appelle River and Wascana Creek
Solar powered autosamplers were programed to collect weekly composite samples
approximately from May to August/September from sample sites along the Qu'Appelle River and
Wascana Creek. In 2008 three autosamplers were installed at sites 4, 5, and 6 (figure 1), in 2009 four
autosamplers were installed in locations based on results obtained in 2008 at sites 1, 4, 5 and 6. In 2010
weekly grab samples were collected at two sites along the Wascana Creek, upstream (50°23'48.71"N
104°29'9.96"W) and downstream (50°29'3.07"N 104°46'42.13"W) of the Regina wastewater treatment
3.0 Results and Discussion
3.1 Conventional End Point PCR assays for detection of host specific Bacteroidales
Saskatchewan water sources.
The host specificity data are summarized in table 2. The Bacteroides
16S rRNA marker was present among all the faecal and raw sewage samples tested (Table 2). The
detection ability of the PCR assay with the human-specific primer (HF183) was 100% on raw sewage
versus 94% (51/54) on human faeces. Similarly, a high detection ability of 98% (119/121) and 100%
(50/50) was obtained for the ruminant (CF128) and pig (PF163)-specific primer, respectively. Notably,
the CF128 marker was unable to distinguish between domestic (cow) and wild (moose, deer, caribou,
bison and wild goat) ruminants confirming that it is a ruminant marker. The HF183 and PF163 primers
both exhibited a specificity of 100% suggesting that no cross-amplification with faecal DNA from the
other species was observed. By comparison, the specificity of the CF128 primer to differentiate
ruminant from the non-ruminant species was 92.8 % (141/152) as the CF128 marker was also detected
in 22% (11/50) of the pig faeces. The fact that this primer failed to differentiate between ruminant and
pig (a non-ruminant) faeces was previously reported in Europe (Gawler et al., 2007; Gourmelon et al.,
2007). Therefore, detection of the CF128 marker alone indicates possible contamination with ruminant
and/or pig faeces. But as appropriately noted by Gourmelon et al. (2007), faecal pollution by pigs can
be confirmed by detection of both CF128 and PF163 markers. In addition, the CF128 marker was
unable to distinguish between domestic and wild ruminants, and consequently its presence in water
does not necessarily indicate faecal contamination from cattle origin. The marker can still be useful in
determining cattle contributions to faecal pollution when land use survey data is also considered with
the CF128 results.
The detection thresholds of the Bacteroidales
markers, indicated the assays provide a high level
of sensitivity. The detection threshold of the Bac32 marker was identical in all the various sources of
faecal material (i.e. detection was possible with a 1 x 10-7 dilution of the fecal sample), whereas it was
two orders of magnitude lower in sewage samples (i.e. 1 x 10 -5). Except for the human faeces, the
detection threshold of each species-specific Bacteroidales PCR assay was reproducible from one
sample to another, and was usually one order of magnitude lower than that obtained for Bac32. In fact,
the CF128 and PF163 markers were both detected up to dilution 1 x 10-6 g of each cow and pig fresh
faeces, respectively. The HF183 marker was detected up to a dilution of 1 x 10 -4 in all the raw sewage
samples, but its detection threshold varied from 1 x 10 -75to 1 x 10 -7 depending on the human faecal
sample tested. This variance in detection threshold, whereas the threshold of detection of the general
Bacteroidales marker Bac32 was similar in the three human faeces, suggests that the prevalence of the
Bacteroidales species carrying the HF183 marker may vary between individuals.
By comparison, detection of E. coli
was up to 10-fold more sensitive than detection of the HF183
in raw sewage. Two previous studies also reported the lower sensitivity of HF183 for detecting sewage
than faecal coliforms (Bernhard and Field, 2000a; Bower et al., 2005). A possible explanation would be
the uneven distribution and/or degradation of target DNA following the death of most bacteria in the
Bacteroidales group due to their lower oxygen tolerance compared to faecal coliforms once released
into sewage (Avelar et al., 1998; Savichtcheva and Okabe, 2006) The PF163-specific PCR assay was
found consistently as sensitive as E. coli
in the pig faecal samples tested. Finally, the sensitivity of the
CF128 marker was 100-fold higher than that of E. coli
, which was detected at levels of ca. 104 CFU g-1
cow faeces. The conventional end-point PCR assays allow sensitive detection of human, ruminant or
pig faecal pollution. The results demonstrated that 1 g of pig or cow faeces could still be detected after
dilution in one cubic meter of water using the PF163 and CF128 specific primer, respectively.
Based on the successful laboratory results detection of the markers was tested in the Qu'Appelle
sampling sites (Figure 1). All water samples tested positive for the general Bac32 Bacteroidales
marker, suggesting that all sites experience some form of faecal pollution. The inability of
Bacteroidales to survive under aerobic conditions, especially at high temperatures encountered during
summer (Avelar et al., 1998; Seurinck et al., 2005), combined with the recurrent detection of the Bac32
marker suggests that all sampling sites were regularly subjected to contamination by faecal material.
However, further research is required to determine the persistence of this marker in environmental
locations such as river sediments. In addition, the fact that indigenous uncultured environmental
Bacteroidales populations could carry this marker should be considered (Lee et al., 2008).
Quantification of the conventional indicator E. coli
suggests that the degree of exposure to the sources
of faecal pollution strongly differed among the sampling locations (Table 3). The detection of E. coli
levels above the recommended limit of 100 CFU per 100 mL (Jones and Shortt, 2005) ranged from
100% in water samples collected from site 4 to 0% in those from the C site (Table 3). On multiple
occasions, the host origin of a portion of the Bacteroides
populations could be determined by using the
16s rRNA markers. Whereas the pig-specific Bacteroides
marker was not detected in any of the water samples, the HF183 and CF128-specific Bacteroides
rRNA markers were detected in 41% (29/70) and 14% (10/70) of the water samples taken throughout
the Qu'Appelle River on 7 and 4 sampling dates, respectively. Notably, these 2 markers were
commonly detected in water sampled from late July till the end of August. As shown in Table 3, the
HF183 or CF128-positive water samples were widely dispersed among the sampling sites. Lack of
detection of the pig marker is not surprising given the low amount of pig farming activity reported in
the Qu'Appelle Valley, and further supports that the CF128 detection is ruminant based. Detection of
the host-specific Bacteroidales markers was significantly different in water samples according to their
level of E. coli
(P < 0.01, Fisher's exact test). There was a higher prevalence of HF183 and CF128 in
water samples with an E. coli level above 100 MPN 100 mL-1 than in samples where E. coli
100 MPN mL-1 [68% (17/25) versus 38% (17/45)]. Notably, 8 of the total 10 CF128-positive samples
were associated with an E. coli level above 100 MPN 100 mL -1. Both HF183 and CF128 markers were
detected at the Wascana, Lumsden or IRPl sites on August 12th when large spikes in E. coli
3.2 Suitability of Quantitative real-time PCR (qPCR) assays for detection of host specific Bacteroidales
markers in Saskatchewan.
Real-time quantitative PCR methods have been recently developed by several different
international groups and used to quantify bovine and human- specific Bacteroidales markers in water
samples (Layton et al., 2006; Kildare et al., 2007; Okabe et al., 2007; Reischer et al., 2007; Shanks et
al., 2008). This approach may be particularly useful in cases of mixed sources of faecal contamination.
Similar to the conventional end point PCR markers (section 3.1) the qPCR makers all displayed
excellent host specificity and sensitivity (Table 5). The levels of detection of the Hum-Bac and Rum-
Bac markers in individual faecal samples indicates that they occur at a high concentration in individual
faeces and to a lower level for the Bov-Bac specific marker (Figures 2-4 ). Therefore detection of the
human and ruminant markers in water following faecal contamination is likely even when the faecal
material is significantly diluted following entry into the water source. The bovine marker is present at a
lower concentration in cattle faeces relative to the ruminant marker (Figures 2, 3). Therefore detection
of the bovine marker in the environment will be more difficult when the cattle faecal contamination
occurs at lower levels. However the advantage to this marker is that when it is detected, particularly at
high concentration it is likely that a large input of cattle faecal pollution has recently occurred.
Based on the successful laboratory results, detection of the markers was tested in the Qu'Appelle
Valley. The advantage of the qPCR approach compared to the conventional end point PCR approach
(section 3.1) is clearly illustrated by the data provided in the field study. This approach allowed a
quantitative measure of the host-specific marker in 2009 sampling, whereas the conventional PCR
assay completed in the Qu'Appelle Valley in 2008 was only qualitative. The results obtained from the
host-specific marker detection support observations of land use surrounding the sampling sites. For
example, sites 1, 2, 3, and 4 are very close to cattle operations and it was not uncommon to frequently
observe cattle in Wascana Creek near the sampling sites. These sites also had the highest concentrations
of ruminant marker and were significantly higher compared to the other sites which experience less
cattle activity (Figure 5) This provides validation that this marker can be used reliable to identify
ruminant related faecal pollution in sites where land use observations would be less clear. Notably, the
concentration of the bovine maker was not significantly higher in sites 1, 2, 3, 4 relative to the other
sites. The lack of a correlation between detection of the ruminant DNA marker and bovine DNA marker
is further emphasized in figure 6, indicating that the Bacteroidales populations detected by the two
markers in cattle faeces are very different. As stated above detection of the bovine marker may be
limited due to its lower abundance in cattle faeces and short persistence (see section 3.3). Despite the
lack of correlation between the ruminant and bovine markers it is notable that the highest
concentrations of the bovine marker did occur in site 3 which is consistently frequented by cattle
(Figure 5). Further testing and optimization of the bovine marker is required before this marker can be
used with confidence in sites where land use history is not known. Interestingly there is a correlation
between the ruminant marker and E. coli
concentrations in the sites likely impacted by cattle. E. coli
concentrations were highest in sites 2, 3, and 4. An r2 value of 0.39 and Pearson p-value of 0.0034 were
obtained when comparing the E. coli
concentration in site 3 with the ruminant marker concentrations.
This result is a good example of the benefit of coupling detection of the conventional E. coli
with the alternative host-specific indicator. In this instance initial E. coli
enumeration provided a
general sense of a faecal contamination problem while the source detection provided information about
the possible source of the contamination.
QPCR detections of the human specific marker were consistent across all sampling sites and no
particular sites stood out. Due to the consistent low level detection of the human marker we were
concerned about its specificity to detect human faecal inputs in the environment. Subsequently, we
selected a reference site that we were confident was not be impacted by human faecal pollution. We
selected Wascana Lake in Regina for this purpose and 3 sampling sites along the shores of the lake
were used. All 3 sites tested negative for the human maker throughout a sampling regime in 2009. This
result along with the validated host specificity and lack of long term persistence (section 3.3) provides
evidence that detection of the human marker in the Qu'Appelle Valley sites is bona-fide. A consistent
low level of human faecal pollution may not be surprising given land use in the area. Several of the
sampling sites are surrounded by human habitations that are not serviced for sewage and it is plausible
that septic systems are frequently supplying low levels of faecal material into the watershed. These data
suggest that further follow-up research on the contributions of human activities (ie sewage treatment
practices) in the Qu'Appelle Valley study area are warranted.
3.3 Persistence of qPCR markers in aquatic ecosystems.
An important criteria of an effective faecal indicator is a short environmental persistence, to
assure detection and quantification of the indicator relates to recent contamination events and not
historical pollution events. The initial data provided by the 2010 studies on Bacteroidales marker
persistence are encouraging. The data from two field trials illustrate that the host-specific markers do
not exhibit long term persistence in a natural water environment (Figures 9, 10). Notably each host
specific marker tested exhibited slightly different persistence times. The bovine specific marker was
consistently the most susceptible to degradation with a rapid decay rate, whereas the ruminant marker
displayed a slightly longer persistence. In all cases the Bacteroidales markers had a significantly shorter
persistence than the conventional E. coli
marker in the trials and did not persist beyond 12 days. Future
persistence experiments are planned and will attempt to better understand the physical, chemical and
biological factors that impact persistence. For example water temperature may play an important role in
persistence (Balleste and Blanch, 2010) and this will be examined further. Our preliminary data
supports this observation as the mean water temperature for the persistence rates with a steeper decay
rate (Figure 9) was 2 ºC higher than the mean water temperature for the persistence rates with the
slower decay rate (Figure 10).
3.4 Detection of bacterial pathogens Salmonella
and the relation to the faecal
indicators E. coli
and host-specifc Bacteroidales markers.
The capacity to detect of Salmonella
during this project was provided
through an in-kind contribution from the Canada Research Chair support provided to Dr. Chris Yost.
This leveraging of funds successfully increased the scope of the research project. In 2009 extensive
sampling for pathogen presence was undertaken at the Qu'Appelle Valley sampling sites. Both
were detected at several sampling sites, while E. coli
O157:H7 was not
detected in any of the samples. An interesting temporal trend was observed for both Salmonella
presence in the Qu'Appelle samples. Frequency of Salmonella
detection was highest in
the month of August while the frequency of Campylobacter
detection was highest in the months of
May and June (Figure 11). A clear explanation is not readily apparent and further investigation will be
required with additional sampling in subsequent seasons to understand better the seasonal association
with pathogen presence in the Qu'Appelle River and Wascana Creek. Notably, the presence of
did not correlate with levels on any faecal indicators (figure 12), a test with Salmonella
was not performed due to the low number of Salmonella
positive samples in the data set. In the
previous year, 2008, there was some weak correlation between the conventional PCR ruminant maker
and the presence of Salmonella
(Fremaux et al. 2009b). However this trend was not observed in 2009.
The persistence and transport of pathogens may be different than the indicators and therefore further
research will be required to determine the ability of the faecal indicators to serve as a proxy for
3.5 Detection of antibiotics in aquatic ecosystems as a tool for assessing water quality.
Autosamplers were positioned at various sampling sites in the Qu'Appelle study in years 2008
(sites 4, 5, and 6) and 2009 (sites 1, 4, 5, and 6). In 2010 autosamplers were not installed, instead
weekly grab samples were obtained at a location downstream of the Regina wastewater treatment plant
and upstream of the plant along Wascana Creek. At the start of this project there was no data regarding
presence of antibiotic residues in the Qu'Appelle River or Wascana Creek. Several antibiotics were
selected for detection based on their use in livestock production and included lincomycin,
erythromycin, sulfamethaxine, iso-chloro-tetracycline, tylosin and monensin. Notably only
erythromycin and monensin were detected during the sampling periods. The optimized methodology
and appropriate controls indicate that lack of detection of the other antibiotics was not due to
experimental error. Simply, these antibiotics were not present in the environmental samples at
concentrations high enough for detection. Conversely, erythromycin was consistently detected while
monensin was infrequently detected in comparison. In 2008, interesting trends were observed across
the sampled sites, where site 4 was clearly distinguished from sites 5, and 6 for presence of both
erythromycin and monensin. In 2009 a new sampling site was introduced on the Wascana Creek, based
on the results from 2008 we targeted an additional sampling site on Wascana Creek. The trend observed
in 2008 was not maintained in 2009 as demonstrated in figure 13. Figure 14 indicates that
concentrations of the antibiotics rose during June through the beginning of August (Figure 14). An
explanation for this is not readily apparent. Future data analysis will investigate hydrological
parameters. In both 2008 and 2009 sites 1 and 4 were consistently the highest for erythromycin and in
fact a strong correlation in erythromycin concentration was observed between the sites, suggesting that
concentrations observed at site 1 may influence those observed at site 4 (Figure 15). Site 1 is in close
proximity to the Regina wastewater treatment plant and maybe be influenced by the plant's effluent. In
2010 samples were taken from Wascana Creek upstream of the wastewater treatment plant and
downstream of the plant. Figure 16 indicates that in 2010 a significantly higher concentration of
erythromycin was observed in water impacted by wastewater effluent. There was no significant
difference between monensin concentrations at the two sites. The data suggests that erythromycin may
be a useful marker for contributions of wastewater discharge. These encouraging results merit further
investigations to determine prevalence of erythromycin in other wastewater effluents in Saskatchewan
as well as to continue monitoring erythromycin concentrations over successive years in Wascana Creek
and the Qu'Appelle River, to better understand the concentration fluxes observed between 2008 and
2009. Persistence of eythromycin in aquatic ecosystems is also an area of future study. In conclusion
the data supports continued investigations on the use of erythromycin as a potential marker for
anthropogenic inputs affecting water quality. Conversely, he data for monensin detection does not
support that monensin detection would be a useful discriminatory water quality marker.
4.0 Conclusions and Recommendations
The host specific Bacteroidales markers tested in this study are robust DNA markers capable of
sensitive and specific detection of faecal material within a Saskatchewan context. Therefore these
markers can be effective at determining both agricultural and human related host-specific sources of
faecal contamination within Saskatchewan water sources. These results are in agreement with other
recent studies which report that the host specificities of the markers are geographically robust in
Europe (Gawler et al., 2007; Gourmelon et al., 2007), Australia (Ahmed et al., 2008), United States
(Shanks et al. 2010a), and Asia (Ahmed et al., 2010). The bovine specific marker tested in this study,
although specific, is not present at high concentrations in cattle faeces and decays rapidly in the
environment. These qualities make it difficult to consistently detect this marker in the environment
since its presence is frequently very close to the detection limit of the protocol. Subsequently a
recommendation for future study is to investigate other bovine specific markers that have been reported
in the literature in the past year (Shanks et al. 2010b). It would also be useful to collect data to test the
supposition that the bovine marker can be effective at identifying severe, recent cattle faecal pollution
events by testing for the marker in environments were intensive livestock activity occurs (ILOs).
The persistence data provided by this study further validates the appeal of the Bacteroidales
based faecal indicator markers. An ideal marker should not persistent in the environment and the
preliminary data from this study suggests these markers, unlike E. coli
, tend to drop below detectable
levels <12 days following a contamination event. Although the preliminary data is encouraging further
persistence studies are required that include persistence in sediments as well as overwintering
persistence. These studies are recommended as future areas of research. Studying persistence of the
markers in relation to pathogen persistence is also recommended as a future area of study given that
marker presence did not provide a direct linkage to pathogen presence. The causes for notable seasonal
abundance of both Salmonella
(summer/August) and Campylobacter
(spring) also merit further
The methodology developed to detect antibiotics related to human and agricultural was
successful. Using the developed methodology on sampling of the Qu'Appelle Valley suggests that
antibiotic detection, in this case erythromycin, can be added to a water quality monitoring toolbox in
both urban and agriculturally influenced watersheds. Furthermore the results from this novel approach
highlight the need for expanded research. In particular a more comprehensive analysis of erythromycin
prevalence at various sampling sites through out Saskatchewan and a study on its persistence in aquatic
ecosystems are warranted.
MST approach has been demonstrated to be a reliable indicator of host
specific faecal contributions in a prairie watershed that is important for irrigation of fresh produce. The
data from this study suggests remediation of bovine faecal pollution at particular sites in Wascana
Creek is appropriate to protect the water quality. An important future study would be to re-asses the
source tracking markers following a remediation strategy to illustrate the effectiveness of the host-
specific remediation approach. Ultimately, it is hoped that these types of future studies will showcase
the usefulness of Bacteroides
MST approaches and lead to the incorporation of MST technologies in
future policy development for protecting irrigation water in Saskatchewan and Canada. Our ability to
identify and eliminate faecal contamination in irrigation water, now and in the future, is essential to
ensure that fresh produce remains free of microbial threats (Solomon, 2003; Byrne et al., 2006).
Support from the Saskatchewan Ministry of Agriculture Agriculture Development Fund has
been clearly acknowledged in all public and scientific venues where results from this project were
presented. In particular the publications resulting from this research are in high impact journals. Water
Research is ranked as the top journal in Water Resources by Journal Citation Reports (2010) and has an
impact factor of 4.355. Applied and Environmental Microbiology is the number one cited journal in
Microbiology and has an impact factor of 3.686. Future publications that will acknowledge the ADF
funding will also be targeted to high impact high readership journals.
The results from this work have also been presented at international conferences and workshops
in France thereby further increasing the international exposure of the Saskatchewan Agriculture and
Food Agriculture Development Fund.
The results have also been presented in seminars focused specific for irrigation users. I have
presented at the Saskatchewan Vegetable Growers Association Annual General Meeting on several
occasions and acknowledged the generous support from the Agriculture Development Fund.
Furthermore participating in these seminars has allowed me to transfer research knowledge to end users
who can benefit from the research.
6.0 Literature Cited
Ahmed, W., Yusuf, R., Hasan, I., Goonetilleke, A., Gardner, T. 2010. Quantitative PCR assay of
sewage-associated Bacteroides markers to assess sewage pollution in an urban lake in Dhaka,
Bangladesh. Canadian Journal of Microbiology 56: 838-845.
Ahmed, W. Stewart, J., Powell, D., Gardner, T. 2008. Evaluation of Bacteroides markers for the
detection of human faecal pollution. Letters in Applied Microbiology 46: 237-242.
Avelar, K.E., Moraes, S.R., Pinto, L.J., Silva e Souza, W., G. das, Domingues, R.M., Ferreira,
M.C. 1998. Influence of stress conditions on Bacteroides fragilis
survival and protein profiles,
Zentralbl. Bakteriol 287: 399–409
Balleste, E., Blanch, A.R. 2010. Persistence of Bacteroides species populations in a river as
measured by molecular and culture techniques. Applied and Environmental Microbiology 76:
Bernhard, A.E., Field, K.G. 2000a. A PCR assay to discriminate human and ruminant feces on
the basis of host differences in Bacteroides
genes encoding 16S rRNA. Applied and
Environmental Microbiology 66: 4571–4574.
Bernhard, A.E., Field, K.G. 2000b. Identification of nonpoint sources of fecal pollution in
coastal waters by using host-specific 16S ribosomal DNA genetic markers from fecal anaerobes.
Applied and Environmental Microbiology 66: 1587–1594.
Blanch, A.R., Belanche-Munoz, L., Bonjoch, X., Ebdon, J., Gantzer, C. and 13 others. 2006.
Integrated analysis of established and novel microbial and chemical methods for microbial source tracking. Applied and Environmental Microbiology 72: 5915-5926.
Bower, P.A., Scopel, C.O., Jensen, E.T., Depas, M.M., McLellan, S.L. 2005. Detection of
genetic markers of fecal indicator bacteria in Lake Michigan and determination of their relationship to
densities using standard microbiological methods. Applied and Environmental
Microbiology 71: 8305–8313
Byrne J., Kienzie, S., Johnson, D., Duke, G., Gannon, B., Selinger, B., Thomas J. 2006. Current
and future water issues in the Oldman River Basin of Alberta, Canada. Water Science and Technology 53: 327-334.
Campagnolo, E.R., Johnson, K.R., Karpati, A., Rubin, C.S., Kolpin, D.W., Meyer, M.T. and 5
others. 2002. Antimicrobial residues in animal waste and water resources proximal to large-scale swine and poultry feeding operations. Science of the Total Environment. 299: 89-95.
Centers for Disease Control and Prevention. 2008. Outbreak of Salmonella
infections associated with multiple raw produce items-United States. MMWR 2008: 57: 929-934
Cessna A.J. Larney, F.J., Kerr, L.A., Bullock, M.S. 2006. Transport of triflualin on wind-eroded
sediment. Canadian Journal of Soil Science. 86: 545-554.
Field, K.G., Samadpour, M. 2007. Fecal source tracking, the indicator paradigm, and managing
water quality, Water Research 41: 3517–3538
Fiksdal, L., Maki, J.S., LaCroix, S.J., Staley, J.T. 1985. Survival and detection of Bacteroides
spp., prospective indicator bacteria. Applied and Environmental Microbiology 49: 148–150
Forrest, F. et al. 2006. Livestock pharmaceuticals in agricultural streams: a scoping study for
Alberta. Alberta Agriculture, Food and Rural Development, Edmonton, Alberta.
Fremaux, B., Boa, T., Chaykowski, A., Kasichayanuia, S., Gritzfeld, J., Braul, L., Yost, C. 2009a.
Assessment of the microbial quality of irrigation water in a prairie watershed. Journal of Applied Microbiology. 106: 442-454.
Fremaux, B. Gritzfeld, J., Boa,, T., Yost, C.K. Evaluation of host-specific Bacteroidales 16S
rRNA gene markers as a complementary tool for detecting fecal pollution in a prairie watershed. Water Research 43: 4838-4849.
Gawler, A.H., Beecher, J.E., Brandao, J. Carroll, N.M., Falcao, L., Gourmelon, M. and 8
others. 2007. Validation of host-specific Bacteriodales
16S rRNA genes as markers to determine the
origin of faecal pollution in Atlantic Rim countries of the European Union, Water Research 41: 3780–
Gourmelon, M., Caprais, M.P. Segura, R., Le Mennec, C., Lozach, S., Piriou, J.Y., Rince, A.
Evaluation of two library-independent microbial source tracking methods to identify sources of fecal
contamination in French estuaries Applied and Environmental Microbiology
Hall, R.I., Leavitt, P.R., Quinlan, R., Dixit, A.S. and Smol, J.P. 1999 Effects of agriculture,
urbanization, and climate on water quality in the northern Great Plains. Limnology and Oceanography 44: 739–756.
Hirsch R., Ternes, T., Haberer, K., Kratz, K.L. 1999. Occurrence of antibiotics in the aquatic
environment. Science of the Total Environment 225: 109-118.
Ishii, S., Sadowsky, M.J. 2008. Escherichia coli
in the environment: Implications for water
quality and human health.
Jones, S. and Shortt, R. 2005. Improving on-farm food safety through good irrigation practices.
OMAFRA Factsheet 05–059.
Kildare, B.J., Leutenegger, C.M., McSwain, B.S., Bambic, D.G., Rajal, V.B. Wuertz, S. 2007.
16S rRNA-based assays for quantitative detection of universal, human-, cow-, and dog-specific fecal
Bacteroidales: a Bayesian approach. Water Research 41: 3701–3715.
Kuchta, S.L., Cessna, A.J. 2009. Fate of lincomycin in snowmelt runoff from manure-amended
pasture. Chemosphere 76: 439-446.
Layton, A., McKay, L., Williams, D., Garrett, V., Gentry, R., Sayler, G. 2006. Development of
16S rRNA gene TaqMan-based real-time PCR assays for estimation of total, human, and
bovine fecal pollution in water. Applied and Environmental Microbiology 72: 4214–4224
Lee, Y.J., Molina, M., Santo Domingo, J.W., Willis, J.D., Cyterski, M., Endale, D.M., Shanks,
O.C. 2008. Temporal assessment of the impact of exposure to cow feces in two watersheds by multiple
host-specific PCR assays. Applied and Environmental Microbiology 74: 6839–6847
Okabe, S., Okayama, N., Savichtcheva, O. Ito, T. 2007. Quantification of host-specific
16S rRNA genetic markers for assessment of fecal pollution in freshwater.
Applied Microbiology and Biotechnology 4: 890–901.
Pomeroy, J.W., De Boer, D. and Martz, L.W. 2005. Hydrology and water resources of
Saskatchewan. Centre for Hydrology Report, Centre for Hydrology, University of Saskatchewan, Saskatoon, pp. 10–11.
Reischer, G.H., Kasper, D.C., Steinborn, R., Farnleitner, A.H. Mach, R.L. 2007. A quantitative
real-time PCR assay for the highly sensitive and specific detection of human faecal influence in spring
water from a large alpine catchment area. Letters in Applied Microbiology 44: 351–356
Santo Domingo, J.W., Bambic, W., Dustin, G., Edge, T.A. Wuertz, S. 2007. Quo vadis source
tracking? Towards a strategic framework for environmental monitoring of fecal pollution. Water
Research 41: 3539-3552
Savichtcheva, O., Okabe, S. 2006. Alternative indicators of fecal pollution: relations with
pathogens and conventional indicators, current methodologies for direct pathogen monitoring and
future application perspectives. Water Research 40: 2463–2476
Seurinck, S. Defoirdt, T., Verstraete, W., Siciliano, S.D. 2005. Detection and quantification of
the human-specific HF183 Bacteroides
16S rRNA genetic marker with real-time PCR for assessment of
human faecal pollution in freshwater. Environmental Microbiology 7:249-259
Sivapalasingam S., Friedman, C.R., Cohen, L., Tauxe, R.V. 2004. Fresh produce: A growing
cause of outbreaks of foodborne illness in the United States, 1973 through 1997. Journal of Food Protection 67: 2342-2353
Steele, M., Odumeru, J. 2004. Irrigation water as source of foodborne pathogens on fruit and
vegetables. Journal of Food Protection 67: 2839-2849
Shanks, O.C., White, K., Kelty, C.A., Sivaganesan, M., Blannon, J., and 3 others. 2010a.
Performance of PCR-Based assays targeting Bacteroidales genetic markers of human fecal pollution in sewage and fecal samples. Environmental Science and Technology 44: 6281-6288.
Shanks, O.C., White, K., Kelty, C.A., Hayes, S., Sivaganesan, M., Jenkins, M., Varma, M.,
Haugland, R.A. 2010b. Performance assessment PCR-based assays targeting Bacteroidales genetic markers of bovine fecal pollution.
Shanks, O.C., Atikovic, E., Blackwood, A.D., Lu, J.R., Nobel, R.T., Domingo, J.S. and 3 others.
2009. Quantitative PCR for detection and enumeration of genetic markers of bovine fecal pollution. Applied and Environmental Microbiology 74: 745-752.
Shuval, H. 2007 Evaluating the world new health organization's 2006 health guidelines for
wastewater. Wastewater reuse – risk assessment. In Decision-Making and Environmental Security ed. pp. 279–287. Netherlands: Springer
Solomon, E.B., Pang, H.J., Matthews, K.R. 2003. Persistence of Escherichia coli
lettuce plants following spray irrigation with contaminated water. Journal of Food Protection 66: 2198-2202.
van Elsas, J.D., Semenov, A.V., Costa, R., Trevors, J.T. 2011. Survival of Escherichia coli
environment: fundamental and public health aspects. The ISME Journal 5: 173-183.
7.1 Papers Published:
*Fremaux, B., Boa, T., Yost, C.K. 2010. Quantitative real-time PCR assays for sensitive
detection of Canada Goose-specific faecal pollution in water sources. Applied and Environmental
Microbiology. 76: 4886-4889.
*although this paper was not a direct deliverable in the ADF grant the knowledge generated
during the ADF work greatly facillitated the completion of development of a goose marker and
consequently the support of ADF was acknowledged.
Fremaux, B., Gritzfeld, J., Boa, T., Yost, C.K. 2009. Evaluation of host-specific Bacteroidales
16S rRNA gene markers as a complementary tool for detecting faecal pollution in a prairie watershed.
Water Research. 43: 4838-4849
7.2 Papers in preperation:
Fremaux, B., Boa, T., Yost, C.K. Quantitative detection of host-specific Bacteroidales gene
markers and their persistence in a prairie watershed. to be submitted to Environmental Science and
Technology impact factor 4.630
Yost, C.K., Baily, J., Boat, T., Cessena, A. Detection of antibiotics in a mixed agriculture and
urban prairie watershed and use as an indicator of water quality. to be submitted to Water Air and Soil
Pollution impact factor 1.676
7.3 Invited Talks:
Yost, C.K. Protecting water from microbial threats through the development of new
faecal pollution indicators. Department of Biology Winter Seminar Series, University of
Regina, Regina, SK. January 14, 2011.
Yost, C.K., Fremaux, B., Boa T. The development and application of microbial
source tracking technologies in Canada. conference. State of Progress on Microbial Source
Tracking (MST) for the Identification of Faecal Pollution. Brest, France. October 29, 2010.
Yost, C.K., Fremaux, B., Boa T. Protecting irrigation water from microbial threats.
Saskatchewan Vegetable Growers Annual Conference. Saskatoon, Saskatchewan. November
Freamux, B., Yost, C.K. Detecting sources of faecal contamination in a
Saskatchewan watershed. Franco-Canadian Workshop on methods to detect and assess faecal
pollution in the environment. University of Lyon, Lyon France. November 1-4, 2009.
Yost, C.K., Fremaux, B., Gritzfeld, J. Ensuring food safety by improving the
microbial quality of irrigation water: a study of the Qu'Appelle River in Southern
Saskatchewan. 6th International Symposium on Public Health and the Agricultural Rural
Ecosystem. Saskatoon, SK. October 19-23, 2008.
Yost, C.K., Fremaux, B., Gritzfeld, J. Protecting the safety of fresh produce by
improving the microbial quality of irrigation water: A study of the Qu'Appelle River in
Southern Saskatchewan. 58th Annual Meeting for the Canadian Society of Microbiologists.
University of Calgary, Calgary, AB. June 9-12, 2008.
Yost, C.K, Fremaux, B., Boa, T., Baily, J., and Cessna, A. Assessing water quality in
agriculturally influenced watersheds using Bacteroidales
-based microbial source tracking
and detection of antibiotics. 59th Annual Conference of the Canadian Society of
Microbiologists. Concordia University, Montreal, Quebec. June 15-18, 2009.
Fremaux, B., Gritzfeld, J.G., Boa, T.B., Chaykowski, A.C., Kasichayanula, S.K.,
Braul, L.B., and Yost, C.K. Improving the microbial quality of irrigation water in
Saskatchewan using monitoring and microbial source tracking. 58th Annual Meeting for the
Canadian Society of Microbiologists. University of Calgary, Calgary, AB. June 9-12, 2008.
I participated in a workshop held at the University of Lyon in Lyon France November 1-4,
2009. The purpose of the workshop was to discuss research approaches on detecting and
assessing faecal pollution in the the environment and to formulate new collaborations with
French researchers and my lab.
A workshop was hosted by the Yost lab and held at the University of Regina on September
21-22, 2010. The workshop brought together Canadian researchers who are experts on
microbial source tracking and included both academia and government agencies. In addition
a collaborator from IFREMER France presented the keynote, this collaboration was formed
following the workshop in Lyon, France, 2009. The program is presented below.
11:30 am Arrival and Check-in
12:00 pm Lunch and workshop overview
1:00 pm Keynote speaker: Michèle Gourmelon, IFREMER, France: "Development of Bacteroidales markers to identify the origin of the faecal pollution in coastal areas in France"
Different Source tracking markers
2:00 pm Richard Villemur, INRS-Institut Armand-Frappier: Mitochondrial DNA as a Tracer for Faecal Contamination Source Tracking. A case study: The L'Assomption river Watershed
2:30 pm Norman Neumann, University of Alberta/Alberta Health Services: "Source tracking faecal pollution in water using parasites and viruses"
3:00 pm Tom Edge, Environment Canada: "From library-dependent to library-independent MST approaches in studying beach closures around the Great Lakes"
3:30 pm Luke Mason, National Research Council: Technical approaches in Mitochondrial Source Tracking
4:00 pm return to hotel
6:00 pm dinner and discussions
Examples of current Source Tracking Approaches with reference to pathogens
9:30 am Kari Dunfield, University of Guelph: "Effect of tillage and timing of liquid manure application on the transport of N, P and bacteria to surface and subsurface waters"
10:00 am Vic Gannon, Public Health Agency of Canada- "Molecular subtypes of Campylobacter
spp., Salmonella enteric
a, and Escherichia coli
O157:H7 isolated from faecal and surface water samples in the Oldman River watershed, Alberta, Canada"
11:00 am Chris Yost, University of Regina: "Detection of Bacteroidales based markers in the Qu'Appelle Valley watershed"
11:30 am Rob Jamieson, Dalhousie University: "Bacterial transport and microbial source tracking investigations within the Thomas Brook Watershed"
1:00-4:00 pm Discussion on implementation of collaborative projects.
State of Science and Readiness for Widespread Application, some suggested points for
1. What are the necessary research endeavours required to move forward MST to wider application and into policy development?2. Opportunities for inter-lab method optimization/validation3. Which MST markers represent suitable options for an inter-lab study4. Possible funding sources to pursue for collaborations5. other points to increase implementation of MST in policy development?-- ie linking source tracking to agricultural BMP (best management practices) studies?
4:00 pm closing remarks and conclusion of meeting.
Table 1: PCR primer sequences for host specific Bacteroidales 16S rRNA gene markers used for
conventional PCR assays.
Anneal ºC Amplicon Size Reference
Bernhard and Field (2000b)
Bernhard and Field
species 16S rRNA gene
HF183F ATCATGAGTTCACATGTCCG Human Bacteroides
Bernhard and Field
species 16S rRNA gene
GCGGATTAATACCGTATGA Pig Bacteroides
Dick et al. (2005)
Table 2: Detection of the host-specific Bacteroides
markers in different species present in
Saskatchewan, Canada using conventional PCR.
Total human samples
Deer, white tailed
Total animal samples
Table 3: Levels of E. coli,
and incidence of the host-specific Bacteroides
16s rRNA markers and
bacterial enteric pathogens in the different sites sampled weekly along the Qu'Appelle River between
June and September 2008
Min-Max E. coli
Mean E. coli
Incidence of Incidence of
MPN= most probable number. We did not detect the presence of the swine specific marker in any of the
samples and therefore it is not included on the table.
a Site numbers correspond to figure 1
b The mean total amount of E. coli
is calculated from the addition of the values obtained during the
irrigation season (June-August)
c n, number of samples with an E. coli
level > 100 MPN 100 mL-1; N, number of total samples tested
for the corresponding period
Table 4: PCR primer and probe sequences used to detect host specific Bacteroidales
using a Quantitative PCR approach.
Layton et al. (2007)
Reicher et al. (2007)
Reischer et al. (2006)
Shanks et al. (2008)
Table 5: Detection of the host-specific Bacteroides
markers in different species present in
Saskatchewan, Canada using the quantitative PCR markers.
Detection of the Bacteroidales
Detection of the
markers via QPCR
Deer, white tailed
Table 6: Levels of E. coli
and the Bacteroidales
16S rRNA markers in the different sites sampled weekly along the Qu'Appelle River during the studying
(MPN 100 ml-1)
(n/N) (%) number 100 ml-1) (n/N) (%) number 100 ml-1) (n/N) (%) number 100 ml-1) (n/N) (%) number 100 ml-1) (n/N) (%)
Location of sampling sites along the Qu'Appelle River and Wascana Creek.
Aerial view of the different sampling sites (from site 1 to site 6 and site C) along the Qu'Appelle River.
In 2008 a site within Lumsden was also sampled, indicated by "L"The arrows indicate the direction of
flow. Potential anthropogenic inputs are also indicated on the figure (Aerial photograph courtesy of
). The treatment plant is the City of Regina sewage treatment plant.
Figure 2. Box plot representation of concentration of Ruminant qPCR marker (Rum-Bac) in individual
cattle faecal samples.
Figure 3. Box plot representation of concentration of Bovine specific qPCR marker (Bov-Bac) in
individual cattle faecal samples
Figure 4. Box plot representation of concentration of Human specific qPCR marker (Hum-Bac) in
individual human faecal samples
Figure 5. Quantitative detection of the ruminant and bovine specific markers using qPCR during the
2009 sampling, May-September.
Top panel is the ruminant marker, bottom panel is the bovine marker. Sites 1 to 4 are similar and
distinct from sites C, 5 and 6 in the ruminant maker (Kruskal-Wallis, P>0.0001). No differences in
values were observed for the bovine marker (Kruskal-Wallis, P>0.05).
Figure 6. Detection of the ruminant specific maker (Rum-Bac) and its relation to detection of the
bovine specific marker (Bov-Bac) during the 2009 sampling season.
The regression analysis indicates no correlation between the two markers.
Figure 7. Quantitative detection of the human specific marker using qPCR during the 2009 sampling,
No significant difference in mean concentrations (Kruskal-Wallis, P>0.5). 3 reference sites in Wascana
Lake tested negative during weekly sampling in Fall 2009.
Figure 8. Enumeration of E. coli
during the 2009 sampling season using the Collilert enumeration
The red line indicates the value of 100 E. coli
CFU 100 ml-1, which is the maximum allowable limit
suggested for irrigation water use as proposed by the Canadian Council of Ministers of the
Environment. Sites 5 and 6 are close to irrigation water pumps.
Figure 9. Persistence trial and rates of decay of Bacteroidales DNA markers and E. coli
from July 12 to July 26, 2010, in Wascana Creek.
Panel A: red diamond is All-Bac maker, blue square is E. coli
. Panel B: limit of detection is log 2/assay,
the blue square is Hum-Bac marker decay, the yellow triangle is Bov-Bac maker decay, the red
diamond is Rum-Bac marker decay
Figure 10. Persistence trial and rates of decay of Bacteroidales DNA markers and E. coli
from August 9 to August 24, 2010, in Wascana Creek.
Panel A: red diamond is All-Bac maker, blue square is E. coli
. Panel B: limit of detection is log 2/assay,
the blue square is Bov-Bac marker decay, the yellow triangle is Hum-Bac maker decay, the red
diamond is Rum-Bac marker decay.
Figure 11. Detection of Campylobacter
in the Qu'Appelle River and Wascana Creek as
a function of month.
Figure 12. Relationship between the Bacteroidales markers and the pathogen Campylobacter
during the 2009 sampling season.
n = 37 n = 52 n = 49
n = 63 n = 37 n = 38
n = 48 n = 50 n = 40
(log copies /100 ml)
(log copies /100 ml)
(log copies /100 ml)
n = 57 n = 47 n = 34
n = 64 n = 74
(log copies /100 ml)
Figure 13. Detection of erythromycin and monensin in the Qu'Appelle River and Wascana Creek
during the 2008 and 2009 sampling seasons.
Concentrations of monensin and erythromycin were significantly different between sampling sites in
2008 (Kruskal Wallis, P < 0.01). In 2009, there was a significant difference between sampling sites
only when testing erythromycin concentrations (Kruskal Wallis, P < 0.001).
Figure 14. Weekly detection of eythromycin during the 2009 sampling season.
Erythromycin Concentration during 2009 sampling
0 4/15/09 4/29/09 5/13/09 6/3/09 6/23/09 7/7/09 7/21/09 8/4/09 8/25/09 9/8/09
Note: Sampling at site 1 did not begin until late May.
Figure 15: Erythromycin concentrations at site 4 maybe influenced by concentrations at site 1.
Figure 16. Erythromycin and monensin concentrations at sites on Wascana Creek upstream and
downstream of the Regina wastewater treatment facility.
D, is the sampling site downstream of the treatment plant while U is the site upstream of the treatment
plant. There was a significant difference between the sampling sites for erythromycin (Kruskal Wallis,
P < 0.001) but not for monensin.
Evaluation of host-specific Bacteroidales 16S rRNA genemarkers as a complementary tool for detecting fecalpollution in a prairie watershed
B. Fremaux, J. Gritzfeld, T. Boa, C.K.
Laboratory for Irrigation Water Protection, Biology Department, University of Regina, 3737 Wascana Parkway, Regina, SK S4S0A2, Canada
Our ability to identify and eliminate fecal contamination of water, now and in the future, is
Received 19 February 2009
essential to reduce incidences of waterborne disease. Bacterial source tracking is a recently
Received in revised form
developed approach for identifying sources of fecal pollution. PCR primers designed by
Bernhard and Field [Bernhard, A.E., Field, K.G., 2000a. A PCR assay to discriminate human
Accepted 22 June 2009
and ruminant feces on the basis of host differences in Bacteroides–Prevotella genes encoding
Published online 27 June 2009
16S rRNA. Appl. Environ. Microbiol. 66(10), 4571–4574] and Dick et al. [Dick, L.K., Bernhard,A.E., Brodeur, T.J., Santo Domingo, J.W., Simpson, J.M., Walters, S.P., Field, K.G., 2005. Host
distributions of uncultivated fecal Bacteroidales bacteria reveal genetic markers for fecal
Bacteroidales 16S rRNA gene markers
source identification. Appl. Environ. Microbiol. 71(6), 3184–3191] for the detection of human
Bacterial source tracking
(HF183), pig (PF163) and ruminant (CF128) specific Bacteroidales 16s rRNA genetic markers
were tested for their suitability in detecting fecal pollution in Saskatchewan, Canada. The
sensitivity and specificity of these primers were assessed by testing eight raw human
sewage samples and 265 feces from 12 different species in Saskatchewan. The specificity of
each primer set was 94%. The accuracy of HF183 and PF163 to distinguish between thedifferent species was 100%, whereas CF128 cross-reacted with 22% of the pig feces.
Occurrence of the host-specific Bacteroidales markers and the conventional indicatorEscherichia coli in relation to several enteropathogens was investigated in 70 water samplescollected from different sites along the Qu'Appelle River (Saskatchewan, Canada). Humanand ruminant fecal markers were identified in 41 and 14% of the water samples, respec-tively, whereas the pig marker was never detected in the river water. The largestconcentrations in E. coli counts were concomitant to the simultaneous detection of HF183and CF128. Thermotolerant Campylobacter spp., Salmonella spp. and Shiga toxin genes (stx1and stx2)-positive E. coli (STEC) were detected in 6, 7 and 63% of the water samples,respectively. However, none of the stx positive water samples were positive for the E. coliO157:H7 gene marker (uidA). Odds ratios analysis suggests that CF128 may be predictive forthe presence of Salmonella spp. in the river investigated. None of the fecal indicators wereable to confidently predict the presence of thermotolerant Campylobacter spp. and STEC.
ª 2009 Elsevier Ltd. All rights reserved.
* Corresponding author. Tel.: þ1 306 585 5223; fax: þ1 306 337 2410.
E-mail address: (C.K. Yost).
0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved.
precisely, host-specific Bacteroidales 16S rRNA genetic markers
were recently identified and PCR primers have been designedfor detection of human (HF183F–Bac708R), ruminant (CF128F–
Fecal pollution of freshwater can have lethal consequences
Bac708R) and swine (PF163F–Bac708R) specific feces (
when enteric pathogens are present in the fecal material
). Several studies suggest that
the host-specificity of the HF183 and PF163 markers may not
public health risks from exposure to fecal dwelling pathogenic
be geographically constrained, whereas the specificity of
microorganisms are substantial, and in addition to the human
CF128 is unresolved in several European countries
tragedy, significant economic loss to a community can occur.
Our ability to identify and eliminate fecal contamination, now
of our knowledge, detection of these markers has not been
and in the future, is essential to reduce incidences of water-
conducted in a Saskatchewan context, and this validation is
borne disease. The low concentration and large diversity of
a critical step before their use for source tracking of fecal
enteric pathogens that can contaminate a watershed preclude
contamination within Saskatchewan. Moreover, further
the direct monitoring of water sources for presence of path-
studies are clearly required to validate an efficient combina-
ogens. Therefore, their potential presence in the aquatic
tion of fecal contamination indicators for predicting the
environment relies mainly on the cultivation and enumera-
presence of pathogens in water and tracing their origin.
tion of fecal indicator bacteria such as total and fecal coli-
The main objective of this study was thus to evaluate the
forms and members of Enterococcus spp. (
suitability of the host-specific Bacteroidales 16s rRNA PCR
). However, some studies have shown that these
primers (HF183F, CF128F and PF163F paired with Bac708R) for
conventional fecal indicators are poorly correlated with the
identifying sources of fecal pollution in freshwater in
presence of pathogenic microorganisms such as Salmonella or
Saskatchewan. The specificity and sensitivity of the PCR
Campylobacter spp., and only indicate the overall pollution of
approach for detection of the human, pig and ruminant-
a water body ). Moreover, detec-
specific Bacteroidales markers were evaluated by testing eight
tion of these indicators in water does not provide information
raw sewage samples and 265 feces originating from 12
regarding the possible sources of the fecal contamination.
different species in Saskatchewan. The PCR detection
Identifying the source of the contamination in water can
thresholds of the different primer sets were assessed and
provide critical information necessary for development of
compared to the standard Colilert-18/Quanti-Tray technology
rational and effective mitigation strategies, such as better
for detecting E. coli in water. The presence of the host-specific
treatment facilities for domestic and industrial effluents and/
Bacteroidales markers and the conventional indicator E. coli in
or management measures for agricultural sources.
relation to several enteropathogens (i.e. Campylobacter coli,
Numerous microbial source tracking (MST) methods are
Campylobacter jejuni, Campylobacter lari, Salmonella spp. and
being developed to help identify sources of fecal pollution in
Shiga toxin genes (stx1, stx2)-positive E. coli (STEC) was
water. These methods can be divided into two groups: culture-
evaluated by analysing freshwater taken from various
based (e.g. antibiotic resistance profile or DNA fingerprinting)
sampling sites along the Qu'Appelle River (Saskatchewan,
and culture-independent (e.g. chemical or DNA amplification
methods) ). Some of the culture-based strategies, such as pulse field gel electrophoresis (PFGE),ribotyping or multiple antibiotic resistance methods, have
Materials and methods
been conducted with Escherichia coli and enterococci. However,they are often labour-intensive requiring the development of
Sampling design and sample collection
a geographically defined host-origin database using signifi-cant numbers of isolates from fecal samples of known origin.
Fecal and raw sewage sample acquisitions
In addition, some researchers question the host-specificity of
To determine the specificity and sensitivity of the different
these indicators and emphasize that they can proliferate and
Bacteroidales primer sets, a total of 265 individual fresh fecal
establish populations in natural environment, which may
samples were aseptically collected from 12 different host
interfere in the tracking of recent contamination events
groups and eight human derived raw sewage samples, which
were obtained from four localities in Saskatchewan (Prince
Certain species from the anaerobic Bacteroidales order have
Albert, Lloydminster, North Battleford and Regina) from
been suggested as alternative biological indicators of fecal
November 2007 to August 2008. Individual human specimens
pollution according to their host-specific distributions, their
(n ¼ 54) were provided by the Disease Control Laboratory of
short survival rate once released into the natural environment
Saskatchewan. Cow (n ¼ 51), chicken (n ¼ 21) and part of the
(due to their low oxygen tolerance) and their abundance in
pig feces (n ¼ 11) were collected as certainly as possible from
feces from warm-blooded animals
separate animals from southern SK farms. Additional fecal
). The considerable advantage
samples from pigs (n ¼ 39) were obtained from the Prairie
of these alternative indicators is their possible application in
Swine Centre (Saskatoon, SK). Fecal samples from other
determining the source of fecal contamination by using
animal species (n ¼ 80) were taken from the Calgary zoo
a culture-independent molecular-based approach, which
(Alberta), except the goose samples (n ¼ 20) that were
produces quick and meaningful results. Among them, PCR
collected from the Wascana Lake (Regina, SK, Canada).
amplification of 16S rDNA from host-specific Bacteroidales has
Samples were transported to the laboratory on ice and kept at
been described as a promising and rapid MST strategy. More
4 C until the time of analysis.
Qu'Appelle River water samples
River. At weekly intervals and for each site, three grab samples
Occurrence of the host-specific Bacteroidales markers and the
(at the middle and each edge of the river) of approximately
conventional indicator Escherichia coli in relation to that of
300 mL were taken from the surface water. The three samples
several enteropathogens was field evaluated in the Qu'Appelle
were used separately for enumeration of E. coli using the
River Basin of southern SK. The Qu'Appelle River extends over
Colilert-18/Quanti-Tray detection system (IDEXX Laborato-
400 km from headwaters near Lake Diefenbaker to join the
ries, Westbrook, Maine, UK) or pooled together for PCR
Assiniboine River in Manitoba. More than 95% of the drainage
detection of the different Bacteroidales markers and bacterial
area is composed of agricultural fields (wheat, barley, canola)
pathogens. In total, 70 river water samples (3 for E. coli
and pastures (especially cattle), but the Qu'Appelle River also
analysis) were collected, transported back to the laboratory on
runs through urbanized areas such as the towns of Lumsden
ice within 4 h of collection and subjected to microbiological
and Craven and is under the influence of effluent from the city
of Regina sewage treatment facility ).
Irrigation of crops in the Qu'Appelle Valley rely on source
Enumeration of E. coli by using the Colilert-18/
water from the Qu'Appelle River which was previously
monitored for its microbial quality (based on enumeration ofE. coli) in our laboratory ). Similar to this
This method was used to process river water samples
earlier study, various sites were sampled weekly from June to
including those used for determining the detection threshold
September 2008; representing the times of most frequent
of the Bacteroidales PCR assays (cf. below). Each undiluted or
irrigation. Sites included the Lumsden town (termed as
appropriately diluted water sample (100 mL) was added to the
Lumsden, ), two sites (termed as Wascana,
Colilert-18 reagent and poured into a 97-well Quanti-Tray
) both located upstream of Lumsden and two locations close
system, sealed and incubated at 37 C for 22 h according to the
to irrigation pumps used to irrigate vegetable crops. The first
manufacturer's instructions (IDEXX). Total coliforms and
one is located immediately downstream of Lumsden (termed
E. coli were simultaneously detected and enumerated by
) and the second downstream of Craven (termed as
counting the number of yellow wells and yellow/fluorescent
) Sampling at the Wascana site was used to
wells under UV light, respectively. The most probable number
determine the contribution of the Wascana creek to fecal
(MPN) could be estimated from a chart provided by the
contamination, one of the major tributaries of the Qu'Appelle
manufacturer. E. coli is a more specific indicator of fecal
Fig. 1 – Aerial view of the different sampling sites ( , Lumsden;
, Qu'Appelle) along the
Qu'Appelle River. The arrows indicate the direction of flow. Potential anthropogenic inputs are also indicated on the figure(Aerial photograph courtesy of ).
used for DNA extraction by using the Power Soil DNA kit (MoBio
therefore only data observed for these bacteria were shown in
Laboratories Inc.). Positive and negative control samples were
our study (results were reported as MPN per 100 mL).
included from the start of each extraction process, throughamplification to electrophoresis. The positive control con-
Sample preparation and DNA extraction for
tained total DNA from the E. coli O157:H7 reference strain 43889
Bacteroidales and bacterial pathogen detection
and from clinical isolates of S. typhimurium (ID 1399EN08), C. lari(ID 2039EN07), C. jejuni (ID 2065EN08) and C. coli (ID 1979EN08)
Analysis of host-specific Bacteroidales
provided by the Saskatchewan Disease Control Laboratory
DNA was directly extracted from 0.25 g of fecal material or
(Regina, SK, Canada).
0.25 mL of raw sewage by using the Power Soil DNA kit (MoBioLaboratories Inc., Carlsbad, CA) according to the manu-
facturer's instructions. Water samples from the Qu'AppelleRiver were first concentrated by filtering 200 mL of water
Detection of host-specific Bacteroidales genetic markers
through a 0.45 mm pore size membrane using a filter funnel
PCR was used to detect host non-specific, human, ruminant
and vacuum system (Millipore Inc., USA). For each sample, the
and pig-specific Bacteroidales 16S rRNA gene markers in fecal
filter was placed in a sterile 15 mL falcon tube (Fisher Scien-
(including raw sewage) and water samples using primer pairs
tific, Ottawa, ON, Canada) containing 5 mL of sterile distilled
previously designed by (
water. The tubes were thoroughly vortexed and 2 mL of this
All amplification reactions were processed in the MyCycler
solution was centrifuged at 11,000 g for 3 min. The pellet
Thermal Cycler (Bio-Rad Laboratories, Hercules, CA), and
was resuspended in 250 ml of sterile distilled water and DNA
carried out in a total volume of 25 ml reaction mixture con-
was extracted from this suspension by using the Power Soil
taining 5 ml of template DNA, 2 mmol l1 MgSO4, 0.2 mmol l1
DNA kit (MoBio Laboratories Inc.) according to the manu-
of each deoxynucleoside triphosphate (Invitrogen, Carlsbad,
USA), 0.4 mmol l1 of each primer (Sigma, Ontario, Canada), 1XTaq reaction buffer, 3 mg of Bovine Serum Albumine (Sigma,
Detection of STEC, E. coli O157:H7, Salmonella spp.
Ontario, Canada) and 1 U of Taq DNA polymerase (UBI, Cal-
and Campylobacter spp
gary, Canada). The amplification program used was as follows:
The method of was employed for detection
5 min initial denaturation of DNA at 95 C, followed by 30
of STEC in water with slight modifications. Briefly, 200 mL of
cycles of 30 s denaturation at 94 C, 30 s primer annealing at
river water from each sampling site were filtered through
the temperature specific for each primer pair and
a 0.45 mm pore size membrane (Millipore) and the filter was
1 min extension at 72 C. Amplification was completed by
placed in a stomacher bag with 125 mL of buffered peptone
a final extension step at 72 C for 6 min. PCR products were
water (BPW) (Sigma, Steinheim, Germany), manually mixed
separated by gel electrophoresis on 1.5% agarose and visual-
and incubated at 37 C overnight. BPW without antibiotic has
ized by ethidium bromide staining.
been recommended by whentesting for stressed or injured STEC cells, as is the case in
Internal amplification control (IAC) for the Human-
environmental samples. To increase the likelihood of detect-
specific Bacteroidales PCR
ing E. coli O157:H7, the most frequently implicated STEC
Although DNA eluates obtained by the Power Soil DNA kit
causing human illness (a 1 mL aliquot of the
(MoBio Laboratories) are of sufficient quality to be directly
BPW culture was transferred to 9 mL of modified trypticase
used for PCR, a synthetic DNA IAC was designed to confirm
soy broth (TSB) (Oxoid, Basingstoke, England) containing
that false–negative results due to Taq inhibitors that could be
20 mg/ml novobiocin (Sigma) and incubated at 42 C for 6 2 h
present in the DNA extract did not occur. The IAC, with the
ment the PCR assay described by was used
to detect the single base-pair mutation on uidA gene (position
þ92) of E. coli O157:H7. Salmonella spp. were selectively
constructed as previously described by . It
enriched by transferring 100 ml of the BPW cultures (above) to
contained the same primer binding sequences as the target
10 mL of Rappaport–Vassiliadis soya peptone broth (RVB)
human-specific Bacteroidales marker DNA (HF183) and was
(Oxoid) followed by incubation at 42 1 C for 24 h. Campylo-
included in each human-specific Bacteroidales PCR reaction. An
bacter spp. were cultured using a modified procedure of
IAC concentration of 109 mM was determined as the most
. Accordingly, 200 mL of each water
appropriate since it does not compete with the target amplifi-
sample were filtered as described above and the membrane
cation. Based on the IAC detection from the HF183 PCR assay,
was incubated in 40 mL of Bolton enrichment broth (BB)
PCR inhibition was not detected in any of the fecal, sewage or
(Oxoid) with selective supplement (Oxoid SR183E) plus 5% (v/v)
water DNA extracts analyzed in the present study.
laked horse blood (Oxoid). Incubation was performed undermicroaerophilic conditions at 42 C for 48 h.
Detection of bacterial pathogens
Following incubation and for each sample, 0.5 mL of each
Individual PCR assays were used to detect the stx genes of
enrichment broth (BPW, TSB, RVB and BB) were mixed together
STEC using degenerate primers that amplify a conserved
(final volume 2 mL) and centrifuged at 11,000 g for 5 min. The
sequence of both stx1 and stx2 genes as described by
supernatant was gently removed and the pellet was resus-
as well as the uidA gene of E. coli O157:H7 (
pended in 250 ml of sterile distilled water. This suspension was
), invA gene of Salmonella spp. ()
Table 1 – Oligonucleotide PCR primers used in this study.
Target microorganism (gene)
Salmonella spp. (invA)
C. jejuni (16S–23S rDNA ITS)
C. lari (16S–23S rDNA ITS)
C. coli (16S–23S rDNA ITS)
E. coli O157:H7 (uidA)
General Bacteroidales (16S rRNA)
Ruminant Bacteroidales species
Human Bacteroidales species
Pig Bacteroidales species (16S rRNA)
and the species-specific genes encoding 16S–23S rDNA
relevant information about the comparison between the E. coli
level and detection by PCR of the host-specific Bacteroidales
Campylobacter species (i.e. C. coli, C. lari and C. jejuni)
in the river water samples. Since C. coli, C. jejuniand C. lari are most often implicated as the causative agent of
Confirmation of PCR amplicons by DNA sequencing
campylobacteriosis (we choose to limitthe detection to these species-specifically. The primer
DNA sequencing was carried out to confirm the identity of the
sequences and expected sizes of amplicons for each PCR assay
PCR fragments obtained for detection of Salmonella spp. (n ¼ 5),
are described in . PCR was performed in 25 ml reaction
C. coli (n ¼ 2), C. jejuni (n ¼ 2) and for randomly selected PCR-
mixtures identical in constituents and volumes to those
amplified sequences (n ¼ 10) from the stx genes primer pair. In
described above for the Bacteroidales PCR assays, with the
addition, human (n ¼ 1) and ruminant (n ¼ 2) specific Bacter-
exceptions of the primers adjusted to a final concentration of
oidales PCR products, obtained from three river water samples
1 mM for the detection of the stx and uidA genes. DNA extracts
associated with high E. coli levels, were subjected to
from both enriched (cf. 2–3–2) or non-enriched (cf. 2–3–1)
sequencing to validate the PCR detection. Briefly, DNA
water samples were separately used as DNA template as an
amplicons to be sequenced were purified from agarose gels
approach to confirm that the targeted pathogenic genes were
using the Gel Extraction Qiaex II kit (Qiagen, Mississauga, ON,
initially present in viable cells. PCR products were verified by
Canada). Purified PCR products generated using host-specific
gel electrophoresis on 1.5% agarose as described above.
Bacteroidales and ITS-specific C. coli primers were cloned usingthe TOPO TA Cloning kit as indicated by the manufacturer
Detection threshold of the different host-specific
Bacteroidales PCR assays
recombinant plasmid DNA from a single clone per TOPOcloning reaction was extracted using a GenElute Plasmid Mini
The detection threshold (SLOD) was determined by using the
Prep kit (Sigma, St Louis, MO). PCR products from all other
method described by , with slight changes.
cases were directly sequenced after purification. Sequencing
Three of the raw sewage, human, cow and pig fecal samples
was performed by the University of Calgary Core DNA
were separately suspended in freshwater taken from the
Sequencing Service (Calgary, Alberta, Canada). Sequence
Qu'Appelle River at a ratio of 1:10. Serial dilutions ranging
similarity searches were carried out using the BLAST algo-
from 101 to 109 were performed for each sample. Each
dilution was subjected to DNA extraction followed by PCR
The DNA sequencing and BLAST analysis using the
analysis for detection of the general Bacteroidales marker Bac32
GenBank database confirmed that the PCR reactions resulted
and the appropriate species-specific Bacteroidales marker as
in appropriate amplification. Overall, sequence similarities
described above. The SLOD was determined as the lowest
were greater than 97% identical between the sequenced
dilution of fecal sample, which gave an amplicon of the
amplicon and the appropriate DNA sequences in the GenBank
expected size. In addition, E. coli were quantified in each
database. The PCR amplicons showing the presence of
dilution by using the Colilert-18 system (IDEXX) to provide
Salmonella (n ¼ 5) exhibited at least 99% similarity with the
invA gene of Salmonella enterica subsp. Enteric. Sequences of the
Table 2 – PCR detection of the host-specific Bacteroidales
PCR-amplified 16S–23S rDNA ITS genes from the positive
markers in different species present in Saskatchewan,
samples for C. jejuni (n ¼ 2) and C. coli (n ¼ 2) had a high degree
of similarity with previously sequenced C. jejuni susbsp. jejuni
(98%) and C. coli (97%) genes, respectively. The sequences of
the stx genes showed up to 97% similarity with stx2d (n ¼ 7 out
of 10 PCR-amplified stx genes), 98% similarity with stx2c (n ¼ 2)
HF183F CF128F PF163F Bac32F
and 99% similarity with stx1 (n ¼ 1). Finally, the Bacteroidales
clones exhibited a high sequence identity (97%) to their
species of origin when subjected to BLAST analysis.
Nucleotide sequence accession numbers
The Bacteroidales 16s rDNA sequences have been submitted to
GenBank under accession numbers FJ716126, FJ716127 and
Deer, white tailed
The sensitivity and specificity of the PCR primer sets were
calculated as: sensitivity ¼ a/(a þ c) and specificity ¼ d/(b þ d ),
where ‘a' is the true positive (positive fecal samples for the
marker of its own species); ‘b' is the false positive (positivefecal samples for the marker of another species; ‘c' is the false
negative (negative fecal samples for the marker of its own
species); ‘d' is the true negative (negative fecal samples for thePCR marker of another species)
ability of 98% (119/121) and 100% (50/50) was obtained for the
The detection of the general and host-specific Bacteroidales
ruminant (CF128) and pig (PF163)-specific primers, respec-
markers in water samples according to their origins (with five
tively. In contrast, the other human (HF134) and ruminant
modalities: IRPc, IRPl, Lumsden, Wascana and Qu'Appelle
(CF193) specific-markers described by
sites) and their level of E. coli (with two modalities: E. coli
were rarely detected among the human (32%, 14/44)
MPN 100 mL1 or E. coli
and cow (16%, 5/32) fecal samples tested, respectively, and
>100 MPN 100 mL1 scored as 0 and 1,
respectively) was analysed, and the statistical significance
were not considered further in this study.
was tested using the Fisher's Exact test (due to the small
The CF128 marker was unable to distinguish between
sample size). The Canadian Council of Ministers of the Envi-
domestic (cow) and wild (moose, deer, caribou, bison and wild
ronment uses 100 MPN 100 mL1 as an indicator for irrigation
goat) ruminants. The HF183 and PF163 primers both exhibited
a specificity of 100% suggesting that no cross-amplification
As previously described by , the odds
with fecal DNA from the other species was observed. By
ratio (with their confidence intervals) were calculated to
comparison, the specificity of the CF128 primer to differen-
assess the likelihood of contamination with thermotolerant
tiate ruminant from the non-ruminant species was 92.8%
Campylobacter spp., Salmonella spp. or STEC according to the
(141/152) as the CF128 marker was also detected in 22% (11/50)
level of E. coli (scored as 0 when it was
of the pig feces.
<100 MPN 100 mL1 and
1 when it was >100 MPN 100 mL1) and the presence of thehost-specific Bacteroidales markers (CF128 and HF183) in
Detection threshold (SLOD) of the Bacteroidales
freshwater samples taken from the Qu'Appelle River. These
statistical calculations were performed using the R softwareversion 2.6.2 ).
The SLOD of the Bac32 marker was identical in all the varioussources of fecal material (i.e. 1 107) (whereas itwas two orders of magnitude lower in sewage samples (i.e.
1 105). Except for the human feces, the SLOD of eachspecies-specific Bacteroidales PCR assay was reproducible from
Host-associated specificity of the Bacteroidales
one sample to another, and was usually one order of magni-
tude lower than that obtained for Bac32. In fact, the CF128 andPF163 markers were both detected up to dilution 1 106 g of
The general Bacteroidales 16S rRNA marker (Bac32) was
each cow and pig fresh feces, respectively The HF183
detected in all the fecal and raw sewage samples tested
marker was detected up to dilution 1 104 in all the raw
(The detection ability of the PCR assay with the
sewage samples, but its SLOD varied from 1 105 to 1 107
human-specific primer (HF183) was 100% for raw sewages
depending on the human fecal sample tested. This lack of
versus 94% (51/54) for human feces. Similarly, a high detection
reproducibility, whereas the SLOD of the general Bacteroidales
Table 3 – Detection thresholds of the different Bacteroidales PCR assays and E. coli in raw sewages and fresh feces fromdiverse origins.
Detection threshold (g feces or mL sewage)
Human feces (HF183)
Raw sewages (HF183)
Cow feces (CF128)
Pig feces (PF163)
a Each Bacteroidales marker tested with fecal or raw sewage samples has been noted between parenthesis in the column ‘origin'.
marker Bac32 was found similar in the three human feces,
IRPl site (In comparison, the Qu'Appelle and IRPc
suggests that the prevalence of the Bacteroidales species
sites experienced minimal contamination with E. coli. A high
carrying the HF183 marker may vary between individuals.
concentration of E. coli was detected on August 12th at the
By comparison, detection of E. coli was up to 10-fold
Wascana, Lumsden and IRPl sites and reached values above
more sensitive than detection of the HF183 in raw sewage
1600 MPN 100 mL1
and up to 100-fold in individual human feces (). ThePF163-specific PCR assay was found consistently as sensitive
PCR detection of the Bacteroidales 16S rRNA
as E. coli in the pig fecal samples tested. Finally, the sensi-
tivity of the CF128 marker was 100-fold higher than that of
The general Bacteroidales 16S rRNA marker was detected in all
E. coli, which was detected at levels of ca. 104 CFU g1 cow
the water samples collected from the different sites. The host-
origin of a portion of the Bacteroidales populations could bedetermined by using the species-specific Bacteroidales 16S
Fecal pollution indicators in the Qu'Appelle River
rRNA markers in 55% (39/70) of the water samples analyzed.
Whereas the pig-specific Bacteroidales 16S rRNA marker was
Enumeration of E. coli using the Colilert-18 system
not detected in any of the water samples, the HF183 and CF128
The Qu'Appelle River exhibited both temporal and spatial
specific Bacteroidales 16S rRNA markers were detected in 41%
variability in contamination with E. coli. The Wascana site
(29/70) and 14% (10/70) of the water samples taken throughout
was the most frequently impacted by high E. coli levels
the Qu'Appelle River. As shown in , the HF183 or CF128-
during the sampling period, followed by the Lumsden and
positive water samples were widely dispersed among the
Table 4 – Levels of E. coli, and incidence of the Bacteroidales 16s rRNA markers and bacterial enteric pathogens in thedifferent sites sampled weekly along the Qu'Appelle River during the studying period.
Samples > Bac32 HF183
CF128 Salmonella Campylobacter
(MPN 100 ml1) (MPN 100 ml1)
6/14 (43) 1/14 (7)
5/14 (36) 3/14 (21)
5/14 (36) 2/14 (14)
6/14 (43) 3/14 (21)
7/14 (50) 1/14 (7)
a The mean total amount of E. coli is calculated from the addition of the values obtained from water samples analyzed during the study period(June–September).
b n, number of water samples with an E. coli level > 100 MPN 100 mL1; N, number of total water samples analyzed.
c The maximum E. coli concentrations occurred simultaneously at the Wascana, Lumsden and IRPl sites on August 12th.
sampling sites, and the Fisher's exact test analysis revealed no
Correlation between E. coli counts, detection of
significant difference between all sites.
species-specific Bacteroidales markers and presence ofbacterial enteric pathogens
PCR detection for the presence of bacterial
Detection of the host-specific Bacteroidales markers was
PCR amplicons could only be detected following enrichment
significantly different in water samples according to their
of the water samples indicating that pathogen loads were very
level of E. coli (P < 0.01, Fisher's exact test) There was
low and that the targeted pathogenic genes were most likely
a higher prevalence of HF183 and CF128 in water samples with
initially present in viable cells. Salmonella and thermotolerant
an E. coli level above 100 MPN 100 mL1 than in the other ones
Campylobacter spp. were detected in seven (10%) out of a total
[68% (17/25) versus 38% (17/45)]. Notably, 80% (8/10) of the
of 70 water samples collected from four different sampling
CF128-positive samples were associated with an E. coli level
sites ). Overall, five (7%) water samples were positive
above 100 MPN 100 mL1. Both HF183 and CF128 markers were
for Salmonella spp. Among the Salmonella spp positive samples,
detected at the Wascana, Lumsden or IRPl sites on August
two sampled from the IRPl site, were also positive for either
12th when large spikes in E. coli numbers and the presence of
C. jejuni or C. coli. A total of four (6%) water samples were
bacterial pathogens also impacted the water.
positive for thermotolerant Campylobacter spp., including two
Among the water samples that tested positive for ther-
for C. jejuni (from the IRPl site) and two for C. coli (from the
motolerant Campylobacter and/or Salmonella spp., six out of
Qu'Appelle and IRPl sites). Notably, samples collected from
seven (86%) also contained an E. coli level above the 100 MPN
the Wascana, Lumsden and IRPl sites on August 12th, the date
100 ml1 limit and/or either the CF128 or HF183 marker, or
of the observed spike in E. coli counts, were positive for either
both. This percentage dropped to 57% (25/44) when consid-
thermotolerant Campylobacter or Salmonella spp., or both. In
ering STEC-positive water samples.
contrast, C. lari and E. coli O157:H7 were not detected in this
Correlation between fecal indicators and the occurrence of
study. Prevalence of STEC in water samples was 63% (44/70).
bacterial pathogens was assessed by calculating the odds
No observable difference in distribution of the STEC-positive
ratios and their 95% confidence intervals (
water samples was obtained among the different sampling
It was not possible to define odds ratio for Salmonella spp.,
thermotolerant Campylobacter spp. and STEC relative to the
Frequency of Bacteroidales markers
Fig. 2 – Detection of Bacteroidales markers (white, unknown origin; shaded black, human origin; black, ruminant origin; shadedgrey, both human and ruminant origins) in water samples according to their level of E. coli.
general Bacteroidales marker Bac32 since this marker was
than fecal coliforms
present in all pathogen-positive samples.
A possible explanation would be the uneven distribution
There was no significant increased probability of detecting
and/or degradation of target DNA following the death of most
any of the bacterial pathogens in samples containing either an
bacteria in the Bacteroidales group due to their lower oxygen
E. coli level above 100 MPN 100 mL1 or the human Bacteroidales
tolerance compared to fecal coliforms once released into
marker HF183, since the confidence interval of the odds ratios
included values 1. Similarly, detection of CF128 showed no
The host-specific Bacteroidales PCR assays were then field
significant association with the occurrence of thermotolerant
tested in the Qu'Appelle River (SK, Canada), a water source
Campylobacter spp. or STEC in water. In contrast, presence of
used routinely for crop irrigation and influenced by several
the CF128 marker was significantly correlated with the pres-
different anthropogenic activities (). Prior monitoring of
ence of Salmonella spp. in water. The odds of detecting
its microbial water quality revealed that levels of E. coli in the
Salmonella spp. when this marker was present were 12 times
river frequently exceed current guidelines for water quality
greater than when it was absent (OR 12.2; 95% CI from 1.7 to
irrigation sources (100 fecal coliform CFU 100 ml1, Environ-
ment Canada, Canadian Water Quality Guidelines) Therefore, identification of the sources of fecalpollution may represent a substantial advance in the resto-
ration and protection of irrigation water quality through theimplementation of appropriate mitigation strategies. Objec-
The first aim of this study was evaluating the host-specificity
tives of this part of the study were to evaluate the suitability of
of selected Bacteroidales 16S rRNA genetic markers using
the Bacteroidales markers to detect fecal contamination in the
a large range of fecal samples (including raw sewages) origi-
Qu'Appelle River, and their occurrence relative to that of
nating from different species in Saskatchewan. The high
sensitivities obtained for the HF183 (100% in raw sewages and
All water samples tested positive for the general Bac32
94% in individual human feces), CF128 (98%) and PF163 (100%)
Bacteroidales marker, suggesting that all sites experience some
markers, indicate that the Bacteroidales host-specificity
form of fecal pollution. The inability of Bacteroidales to survive
detected by this PCR method is present in Saskatchewan. This
under aerobic conditions, especially at high temperatures
finding was comparable to those obtained for both the HF183
encountered during summer
and CF128 markers in regions of Europe or
combined with the recurrent
the United States (Oregon) (), and
detection of the Bac32 marker suggests that all sampling sites
recently in Australia for HF183 ). Similar to
were regularly subjected to contamination by fecal material.
our study, the PF163 marker exhibited a 100% sensitivity when
However, further research is required to determine the
tested in France The Bacteroidales
persistence of this marker in environmental locations such as
species that carry the HF183 or PF163 markers were strictly
river sediments. In addition, the fact that indigenous uncul-
host-specific in our study as no cross-amplification with fecal
tured environmental Bacteroidales populations could carry this
DNA from other species was obtained. Conversely, fecal DNA
marker should be considered Quantification
extracted from 11 pig feces cross-reacted with the CF128
of the conventional indicator E. coli suggests that the degree of
specific primer. The fact that this primer failed to differentiate
exposure to the sources of fecal pollution strongly differed
between ruminant and pig (a non-ruminant) feces was
among the sampling locations. The detection of E. coli levels
previously reported in Europe
above the recommended limit of 100 CFU per 100 mL (
Therefore, detection of the CF128 marker alone
ranged from 100% in water samples collected
indicates possible contamination with ruminant and/or pig
from the Wascana site to 0% in those from the Qu'Appelle site.
feces. But as appropriately noted by ,
The use of host-specific Bacteroidales gene markers
fecal pollution by pigs can be confirmed by detection of both
revealed that the Qu'Appelle River was occasionally impacted
CF128 and PF163 markers. In addition, the CF128 marker was
by human and ruminant fecal contamination as the HF183
unable to distinguish between domestic and wild ruminants,
and CF128 markers were detected in 41 and 14% of the water
and consequently its presence in water does not necessarily
samples analysed, respectively, without significant difference
indicate fecal contamination from agricultural origin.
between the sampling locations. In contrast, the pig-specific
The PCR assays allow sensitive detection of human,
PF163 marker was never detected at any of the sampling sites.
ruminant or pig fecal pollution. The results demonstrated that
This is not surprising given the low amount of pig farming
1 g of pig or cow feces could still be detected after dilution in
activity reported in the Qu'Appelle Valley.
one cubic meter of water using the PF163 and CF128 specific
The HF183 and CF128 markers were preferentially detected
primer, respectively. Notably, detection of the CF128 marker
in water samples associated with a level of E. coli above 100
was found 100-fold more sensitive than detection of E. coli in
MPN 100 ml1. Notably, 80% of the CF128-positive water
the cow feces samples tested. Based on the same protocol,
samples contained a concentration in E. coli above 100 MPN
reported similar CF128-SLOD
100 ml1. The unusually high E. coli concentrations (above
values in cow feces (i.e. 106–107 g feces). Using raw sewages,
1600 MPN 100 mL1) measured on August 12th at the Was-
we demonstrated that the HF183 marker was consistently
cana, Lumsden and IRPl sites, were associated with the
detected up to dilution 1 104 mL sewage, which is up to 10-
simultaneously detection of both the HF183 and CF128
fold less sensitive than E. coli. Two previous studies also
markers. However, the relative contribution of each of these
reported the lower sensitivity of HF183 for detecting sewage
markers to the spikes in E. coli counts could not be assessed
using the qualitative PCR assays. Real-time quantitative PCR
found a significant association between HF183 and
methods have been recently developed by several different
Campylobacter spp., but neither E. coli or the Bacteroidales
international groups and used to quantify bovine and human-
markers were able to predict the presence of Salmonella spp. in
specific Bacteroidales markers in water samples
surface water samples (Oldman River Basin, southern Alberta,
Canada). Based on quantitative data obtained from WWTP and
), which can help to determine the
water surface samples, showed that
various contributions of fecal contamination in cases of mixed
fecal coliforms and the HF183 marker had the greatest
sources. Notably, the Wascana creek continuously receives
predictive value for the occurrence of Salmonella spp. Discrep-
discharges of effluents from the sewage treatment plant
ancies between these studies results from the fact that ther-
(Regina) that may affect downstream water quality in case of
motolerant Campylobacter and Salmonella spp. may be carried
system failure. In addition, the Qu'Appelle River is surrounded
by several host-organisms. As a consequence, the strength in
by numerous livestock operations ), and inappropriate
the relationship between fecal source tracking indicators and
agricultural practices such as direct deposition of cattle waste
presence of these pathogens may vary significantly between
into water could lead to both large E. coli spikes and detection
different watersheds according to the abundance and
of the ruminant-specific marker. The results of this study
frequency of humans and/or animals carrying the pathogens.
emphasize the need for a comprehensive watershed-scaleassessment of fecal pollution in the Qu'Appelle Valleywatershed that should include hydrological studies, and
multiple seasons of data collection. The Bacteroidales genemarkers investigated in this study may provide a valuable tool
– PCR detection of the various host-specific Bacteroidales
to help define the fecal inputs in such a comprehensive study.
markers in fecal samples exhibited a great host-specificity
Salmonella spp., thermotolerant Campylobacter spp. and
and sensitivity. Therefore, it can be a valuable additional
STEC non-O157:H7 were detected in the Qu'Appelle river
tool available to researchers undertaking comprehensive
water during the irrigation season. STEC were detected at high
studies to characterize fecal pollution inputs in the
frequency regardless of the sampling sites and were
frequently found in water relatively unpolluted as indicated
– A statistically significant relationship was demonstrated
by the absence of either conventional or alternative fecal
between occurrence of the ruminant-specific Bacteroidales
indicators. However, it is important to note that not all STEC
marker (CF128) and that of Salmonella spp.
strains carrying the stx genes may be pathogenic to humans.
– Their detection in the Qu'Appelle River emphasizes the
The virulence of STEC is multifactorial, and further evaluation
benefit of combined measurement of both alternative and
of their serotypes and detection of other additional virulence
conventional indicators for a better understanding of the
genes, for example intimin (eae) and enterohemolysin (ehxA)
nature of fecal contamination events.
are required to assess the virulence potential of the STECstrains ). Because ruminants, espe-cially cattle, are considered as the main reservoir of STEC, itwas expected to find a correlation with CF128. The lack ofcorrelation between this marker and STEC may be caused by
the ability of STEC to persist in freshwater (for a period of time much
We thank the two anonymous reviewers for their thoughtful
longer than the CF128 specific Bacteroidales marker. This could
review and helpful comments. The generous support of the
be also due to the high dilution of the cow Bacteroidales marker
Saskatchewan Agriculture, Agriculture Development Fund is
in water which causes problems for their detection. In
greatly appreciated. This work was also supported through
contrast, Salmonella and thermotolerant Campylobacter spp.
the Canada Research Chairs program to C. Yost and the
were detected on few occasions, however detection did occur
Canada–Saskatchewan Water Supply Expansion Program,
at the IRPl site, which is close to the irrigation pump used for
administered by Agriculture and Agri-Food Canada. The
irrigation of vegetable crops.
authors thank Shawn Hansen and the Saskatchewan Vege-
The odds ratios analysis revealed that the CF128 marker
table Growers Association for sampling assistance. The
may predict the presence of Salmonella spp. in the river. This is
authors also acknowledge the many contributors who have
consistent with the fact that Salmonella spp. are found natu-
helped in the collection of fecal samples: Jill Hockaday, Jodene
rally in the intestines of ruminant such as cattle (
Mills and the keepers from the Calgary zoo, the sewage
which are present
treatment plant facilities from Prince Albert, Lloydminster,
in relatively high density in the watershed In contrast,
North Battleford and Regina, Ryan MacDonald from the
none of the Bacteroidales host-specific fecal markers tested in
Disease Control Laboratory of Saskatchewan, the farmers and
this study as well as the conventional indicator E. coli were
the Prairie Swine Centre from Saskatoon.
adequate to predict the presence of thermotolerant Campylo-bacter spp. However, due to the low detection rates of these
pathogens (seven out of the 70 water samples) during theseason, further data is required to confidently assess therelationship of Salmonella and thermotolerant Campylobacter
Ahmed, W., Powell, D., Goonetilleke, A., Gardner, T., 2008.
spp. with such fecal indicators in this study system.
Detection and source identification of faecal pollution in non-
sewered catchment by means of host-specific molecular
library-independent microbial source tracking methods to
markers. Water Sci. Technol 58 (3), 579–586.
identify sources of fecal contamination in French estuaries.
Avelar, K.E., Moraes, S.R., Pinto, L.J., Silva e Souza, W., das, G.,
Appl. Environ. Microbiol. 73 (15), 4857–4866.
Domingues, R.M., Ferreira, M.C., 1998. Influence of stress
Hall, R.I., Leavitt, P.R., Quinlan, R., Dixit, A.S., Smol, J.P., 1999. Effects
conditions on Bacteroides fragilis survival and protein profiles.
of agriculture, urbanization, and climate on water quality in the
Zentralbl. Bakteriol 287 (4), 399–409.
northern Great Plains. Limno. Oceano 44, 739–756.
Bernhard, A.E., Field, K.G., 2000a. A PCR assay to discriminate
Hrudey, S.E., Payment, P., Huck, P.M., Gillham, R.W., Hrudey, E.J.,
human and ruminant feces on the basis of host differences in
2003. A fatal waterborne disease epidemic in Walkerton,
Bacteroides– Prevotella genes encoding 16S rRNA. Appl. Environ.
Ontario: comparison with other waterborne outbreaks in the
Microbiol. 66 (10), 4571–4574.
developed world. Water Sci. Technol 47 (3), 7–14.
Bernhard, A.E., Field, K.G., 2000b. Identification of nonpoint
Hussein, H.S., Bollinger, L.M., 2008. Influence of selective media
sources of fecal pollution in coastal waters by using host-
on successful detection of Shiga toxin-producing Escherichia
specific 16S ribosomal DNA genetic markers from fecal
coli in food, fecal, and environmental samples. Foodborne
anaerobes. Appl. Environ. Microbiol. 66 (4), 1587–1594.
Pathog. Dis 5 (3), 227–244.
Bouvet, J., Bavai, C., Rossel, R., Le Roux, A., Montet, M.P.,
Ihaka, R., Gentleman, R., 1996. R: a language for data analysis and
Ray-Gueniot, S., Mazuy, C., Atrache, V., Vernozy-Rozand, C.,
graphics. J. Comp. Graph. Stat 5, 299–314.
2002. Effects of cutting process on pork meat contamination
Jones, S., Shortt, R., 2005. Improving on-farm food safety through
by verotoxin-producing Escherichia coli (VTEC) and E. coli O157:
good irrigation practices. OMAFRA Factsheet, 05–059.
H7. Int. J. Food Microbiol. 77 (1-2), 91–97.
Johnson, J.Y., Thomas, J.E., Graham, T.A., Townshend, I., Byrne, J.,
Bower, P.A., Scopel, C.O., Jensen, E.T., Depas, M.M., McLellan, S.L.,
Selinger, L.B., Gannon, V.P., 2003. Prevalence of Escherichia coli
2005. Detection of genetic markers of fecal indicator bacteria
O157:H7 and Salmonella spp. in surface waters of southern
in Lake Michigan and determination of their relationship to
Alberta and its relation to manure sources. Can. J. Microbiol.
Escherichia coli densities using standard microbiological
49 (5), 326–335.
methods. Appl. Environ. Microbiol. 71 (12), 8305–8313.
Khaitsa, M.L., Kegode, R.B., Bauer, M.L., Gibbs, P.S., Lardy, G.P.,
Buckalew, D.W., Hartman, L.J., Grimsley, G.A., Martin, A.E.,
Doetkott, D.K., 2007. A longitudinal study of Salmonella
Register, K.M., 2006. A long-term study comparing membrane
shedding and antimicrobial resistance patterns in North
filtration with Colilert defined substrates in detecting fecal
Dakota feedlot cattle. J. Food Prot 70 (2), 476–481.
coliforms and Escherichia coli in natural waters. J. Environ.
Khan, I.U., Edge, T.A., 2007. Development of a novel triplex PCR
Manage 80 (3), 191–197.
assay for the detection and differentiation of thermophilic
Cebula, T.A., Payne, W.L., Feng, P., 1995. Simultaneous
species of Campylobacter using 16S–23S rDNA internal
identification of strains of Escherichia coli serotype O157:H7 and
transcribed spacer (ITS) region. J. Appl. Microbiol. 103 (6),
their Shiga-like toxin type by mismatch amplification mutation
assay-multiplex PCR. J. Clin. Microbiol. 33 (1), 248–250.
Kildare, B.J., Leutenegger, C.M., McSwain, B.S., Bambic, D.G.,
Chiu, C.H., Ou, J.T., 1996. Rapid identification of Salmonella
Rajal, V.B., Wuertz, S., 2007. 16S rRNA-based assays for
serovars in feces by specific detection of virulence genes, invA
quantitative detection of universal, human-, cow-, and dog-
and spvC, by an enrichment broth culture-multiplex PCR
specific fecal Bacteroidales: a Bayesian approach. Water Res.
combination assay. J. Clin. Microbiol. 34 (10), 2619–2622.
41 (16), 3701–3715.
Dick, L.K., Bernhard, A.E., Brodeur, T.J., Santo Domingo, J.W.,
Kreader, C.A., 1998. Persistence of PCR-detectable Bacteroides
Simpson, J.M., Walters, S.P., Field, K.G., 2005. Host
distasonis from human feces in river water. Appl. Environ.
distributions of uncultivated fecal Bacteroidales bacteria reveal
Microbiol. 64 (10), 4103–4105.
genetic markers for fecal source identification. Appl. Environ.
Layton, A., McKay, L., Williams, D., Garrett, V., Gentry, R.,
Microbiol. 71 (6), 3184–3191.
Sayler, G., 2006. Development of Bacteroides 16S rRNA gene
Fey, P.D., Wickert, R.S., Rupp, M.E., Safranek, T.J., Hinrichs, S.H.,
TaqMan-based real-time PCR assays for estimation of total,
2000. Prevalence of non-O157:H7 shiga toxin-producing
human, and bovine fecal pollution in water. Appl. Environ.
Escherichia coli in diarrheal stool samples from Nebraska.
Microbiol. 72 (6), 4214–4224.
Emerg. Infect. Dis 6 (5), 530–533.
Lee, Y.J., Molina, M., Santo Domingo, J.W., Willis, J.D., Cyterski, M.,
Field, K.G., Samadpour, M., 2007. Fecal source tracking, the
Endale, D.M., Shanks, O.C., 2008. Temporal assessment of the
indicator paradigm, and managing water quality. Water Res.
impact of exposure to cow feces in two watersheds by
41 (16), 3517–3538.
multiple host-specific PCR assays. Appl. Environ. Microbiol. 74
Fiksdal, L., Maki, J.S., LaCroix, S.J., Staley, J.T., 1985. Survival and
detection of Bacteroides spp., prospective indicator bacteria.
Milnes, A.S., Stewart, I., Clifton-Hadley, F.A., Davies, R.H.,
Appl. Envrion. Microbiol. 49 (1), 148–150.
Newell, D.G., Sayers, A.R., Cheasty, T., Cassar, C., Ridley, A.,
Fremaux, B., Prigent-Combaret, C., Vernozy-Rozand, C., 2008.
Cook, A.J., Evans, S.J., Teale, C.J., Smith, R.P., McNally, A.,
Long-term survival of Shiga toxin-producing Escherichia coli in
Toszeghy, M., Futter, R., Kay, A., Paiba, G.A., 2008. Intestinal
cattle effluents and environment: an updated review. Vet.
carriage of verocytotoxigenic Escherichia coli O157, Salmonella,
Microbiol. 132 (1-2), 1–18.
thermophilic Campylobacter and Yersinia enterocolitica, in cattle,
Fremaux, B., Boa, T., Chaykowski, A., Kasichayanula, S.,
sheep and pigs at slaughter in Great Britain during 2003.
Gritzfeld, J., Braul, L., Yost, C., 2009. Assessment of the
Epidemiol. Infect 136 (6), 739–751.
microbial quality of irrigation water in a prairie watershed.
Nataro, J.P., Kaper, J.B., 1998. Diarrheagenic Escherichia coli. Clin.
J. Appl. Microbiol. 106 (2), 442–454.
Microbiol. Rev. 11, 142–201.
Gawler, A.H., Beecher, J.E., Brandao, J., Carroll, N.M., Falcao, L.,
O'Reilly, C.E., Bowen, A.B., Perez, N.E., Sarisky, J.P.,
Gourmelon, M., Masterson, B., Nunes, B., Porter, J., Rince, A.,
Shepherd, C.A., Miller, M.D., Hubbard, B.C., Herring, M.,
Rodrigues, R., Thorp, M., Walters, J.M., Meijer, W.G., 2007.
Buchanan, S.D., Fitzgerald, C.C., Hill, V., Arrowood, M.J.,
Validation of host-specific Bacteriodales 16S rRNA genes as
Xiao, L.X., Hoekstra, R.M., Mintz, E.D., Lynch, M.F.Outbreak
markers to determine the origin of faecal pollution in Atlantic Rim
Working Group, 2007. A waterborne outbreak of
countries of the European Union. Water Res. 41 (16), 3780–3784.
gastroenteritis with multiple etiologies among resort island
Gourmelon, M., Caprais, M.P., Segura, R., Le Mennec, C.,
visitors and residents: Ohio, 2004. Clin. Infect. Dis 44 (4),
Lozach, S., Piriou, J.Y., Rince, A., 2007. Evaluation of two
Okabe, S., Okayama, N., Savichtcheva, O., Ito, T., 2007.
Bacteroides 16S rRNA genetic marker with real-time PCR for
Quantification of host-specific Bacteroides–Prevotella 16S rRNA
assessment of human faecal pollution in freshwater. Environ.
genetic markers for assessment of fecal pollution in
Microbiol. 7 (2), 249–259.
freshwater. Appl. Microbiol. Biotechnol 74 (4), 890–901.
Shanks, O.C., Atikovic, E., Blackwood, A.D., Lu, J., Noble, R.T.,
Pangloli, P., Dje, Y., Ahmed, O., Doane, C.A., Oliver, S.P.,
Domingo, J.S., Seifring, S., Sivaganesan, M., Haugland, R.A.,
Draughon, F.A., 2008. Seasonal incidence and molecular
2008. Quantitative PCR for detection and enumeration of
characterization of Salmonella from dairy cows, calves, and
genetic markers of bovine fecal pollution. Appl. Environ.
farm environment. Foodborne Pathog. Dis 5 (1), 87–96.
Microbiol. 74 (3), 745–752.
Read, S.C., Clarke, R.C., Martin, A., De Grandis, S.A., Hii, J.,
Ufnar, J.A., Wang, S.Y., Ufnar, D.F., Ellender, R.D., 2007.
McEwen, S., Gyles, C.L., 1992. Polymerase chain reaction for
Methanobrevibacter ruminantium as an indicator of
detection of verocytotoxigenic Escherichia coli isolated from
domesticated-ruminant fecal pollution in surface waters.
animal and food sources. Mol. Cell. Probes 6 (2), 153–161.
Appl. Environ. Microbiol. 73 (21), 7118–7121.
Reischer, G.H., Kasper, D.C., Steinborn, R., Farnleitner, A.H.,
Walters, S.P., Gannon, V.P., Field, K.G., 2007. Detection of
Mach, R.L., 2007. A quantitative real-time PCR assay for the
Bacteroidales fecal indicators and the zoonotic pathogens E. coli
highly sensitive and specific detection of human faecal
0157:H7, Salmonella, and Campylobacter in river water. Environ.
influence in spring water from a large alpine catchment area.
Sci. Technol 41 (6), 1856–1862.
Lett. Appl. Microbiol. 44 (4), 351–356.
Watterworth, L., Rosa, B., Schraft, H., Topp, E., Leung, K., 2006.
Santo Domingo, J.W., Bambic, D.G., Edge, T.A., Wuertz, S., 2007.
Survival of various ERIC-genotypes of Shiga toxin-producing
Quo vadis source tracking? Towards a strategic framework for
Escherichia coli in well water. Water Air Soil Pollut 177, 367–382.
environmental monitoring of fecal pollution. Water Res. 41
Wilkes, G., Edge, T., Gannon, V., Jokinen, C., Lyautey, E.,
Medeiros, D., Neumann, N., Ruecker, N., Topp, E., Lapen, D.R.,
Savichtcheva, O., Okabe, S., 2006. Alternative indicators of fecal
2009. Seasonal relationships among indicator bacteria,
pollution: relations with pathogens and conventional
pathogenic bacteria, Cryptosporidium oocysts, Giardia cysts,
indicators, current methodologies for direct pathogen
and hydrological indices for surface waters within an
monitoring and future application perspectives. Water Res. 40
agricultural landscape. Water Res. 43 (8), 3615–3628.
Yoder, J., Roberts, V., Craun, G.F., Hill, V., Hicks, L.A.,
Savichtcheva, O., Okayama, N., Okabe, S., 2007. Relationships
Alexander, N.T., Radke, V., Calderon, R.L., Hlavsa, M.C.,
between Bacteroides 16S rRNA genetic markers and presence of
Beach, M.J., Roy, S.L.Centers for Disease Control and
bacterial enteric pathogens and conventional fecal indicators.
Prevention (CDC), 2008. Surveillance for waterborne disease
Water Res. 41 (16), 3615–3628.
and outbreaks associated with drinking water and water not
Seurinck, S., Defoirdt, T., Verstraete, W., Siciliano, S.D., 2005.
intended for drinkingdUnited States, 2005–2006. MMWR
Detection and quantification of the human-specific HF183
Surveill. Summ 57 (9), 39–62.
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, July 2010, p. 4886–4889
Copyright 2010, American Society for Microbiology. All Rights Reserved.
Quantitative Real-Time PCR Assays for Sensitive Detection of Canada
Goose-Specific Fecal Pollution in Water Sources䌤†
B. Fremaux, T. Boa, and C. K. Yost*
Laboratory for Irrigation Water Protection, Biology Department, University of Regina, 3737 Wascana Parkway,
Regina, Saskatchewan S4S 0A2, Canada
Received 15 January 2010/Accepted 20 May 2010
Canada geese (Branta canadensis) are prevalent in North America and may contribute to fecal pollution of
water systems where they congregate. This work provides two novel real-time PCR assays (CGOF1-Bac and
CGOF2-Bac) allowing for the specific and sensitive detection of Bacteroides 16S rRNA gene markers present
within Canada goose feces.
The Canada goose (Branta canadensis
) is a prevalent water-
material). Two hundred eighty-eight clones were randomly se-
fowl species in North America. The population density of Can-
lected and subjected to DNA sequencing (at the Plant Biotech-
ada geese has doubled during the past 15 years, and the pop-
nology Institute DNA Technologies Unit, Saskatoon, SK, Can-
ulation was estimated to be close to 3 million in 2007 (4).
ada). Representative sequences of each operational taxonomic
Canada geese often congregate within urban settings, likely
unit (OTU) were recovered using an approach similar to that
due to available water sources, predator-free grasslands, and
described by Mieszkin et al. (16). Sequences that were less than
readily available food supplied by humans (6). They are sus-
93% similar to 16S rRNA gene sequences from nontarget host
pected to contribute to pollution of aquatic environments due
species in GenBank were used in multiple alignments to identify
to the large amounts of fecal matter that can be transported
regions of DNA sequence that were putatively goose specific.
into the water. This can create a public health threat if the fecal
Subsequently, two TaqMan fluorescent probe sets (targeting
droppings contain pathogenic microorganisms (6, 7, 9, 10, 12,
markers designated CGOF1-Bac
) were de-
13, 19). Therefore, tracking transient fecal pollution of water
signed using the RealTimeDesign software provided by Biosearch
due to fecal inputs from waterfowl, such as Canada geese, is of
Technologies (http://www.biosearchtech.com/). The newly de-
importance for protecting public health.
signed primer and probe set for the CGOF1-Bac
PCR detection of host-specific 16S rRNA gene sequences
CG1F (5⬘-GTAGGCCGTGTTTTAAGTCAGC-3⬘) and CG1R
of fecal origin has been described as a
(5⬘-AGTTCCGCCTGCCTTGTCTA-3⬘) and a TaqMan probe
promising microbial source-tracking (MST) approach due to
its rapidity and high specificity (2, 3). Recently, Lu et al. (15)
ACACTTGAG-Black Hole Quencher 1 [BHQ-1]-3⬘), and the
characterized the fecal microbial community from Canada
assay had primers CG2F (5⬘-ACTCAGGGATAG
geese by constructing a 16S rRNA gene sequence database
CCTTTCGA-3⬘) and CG2R (5⬘-ACCGATGAATCTTTCTTTG
using primers designed to amplify all bacterial 16S rRNA gene
TCTCC-3⬘) and a TaqMan probe (5⬘-FAM-AATACCTGATGC
sequences. The authors reported that the majority of the 16S
rRNA gene sequences obtained were related to Clostridia
for the Canada goose-associated Bacteroides
16S rRNA prim-
and to a lesser degree Bacteroidetes
, which represent
ers were verified through in silico
analysis using BLASTN (1)
possible targets for host-specific source-tracking assays.
and the probe match program of the Ribosomal Database
The main objective of this study was to identify novel Bac-
Project (release 10) (5). Host specificity was further confirmed
16S rRNA gene sequences that are specific to Can-
using DNA extracts from 6 raw human sewage samples from
ada goose feces and design primers and TaqMan fluorescent
various geographical locations in Saskatchewan and 386 fecal
probes for sensitive and specific quantification of Canadagoose fecal contamination in water sources.
samples originating from 17 different animal species in Sas-
Primers 32F and 708R from Bernhard and Field (2) were used
katchewan, including samples from Canada geese (n
to construct a Bacteroidales
-specific 16S rRNA gene clone library
(Table 1). An existing nested PCR assay for detecting Canada
from Canada goose fecal samples (n
⫽ 15) collected from grass
goose feces (15) (targeting genetic marker CG-Prev
lawns surrounding Wascana Lake (Regina, SK, Canada) in May
Table S1 in the supplemental material) was also tested for
2009 (for a detailed protocol, see File S1 in the supplemental
specificity using the individual fecal and raw sewage samples(Table 1). All fecal DNA extracts were obtained from 0.25 g offecal material by using the PowerSoil DNA extraction kit (MoBio Inc., Carlsbad, CA) (File S1 in the supplemental material
* Corresponding author. Mailing address: Laboratory for Irrigation
Water Protection, Biology Department, University of Regina, 3737
provides details on the sample collection).
Wascana Parkway, Regina, SK S4S 0A2, Canada. Phone: (306) 585-
The majority of the Canada goose feces analyzed in this
5223. Fax: (306) 337-2410. E-mail: email@example.com.
study (94%; 95 of 101) carried the Bacteroidales
† Supplemental material for this article may be found at http://aem
genetic marker designated All-Bac
, with a relatively high me-
䌤 Published ahead of print on 28 May 2010.
dian concentration of 8.2 log
copies g⫺1 wet feces (Table 1
ESTIMATION OF CANADA GOOSE FECAL POLLUTION IN WATER
TABLE 1. Specificities of the CGOF1-Bac
with fecal DNA from other host groups, while cross-amplifi-
f5 PCR assays for different species
cation for the CGOF1-Bac
assay was limited to one pigeon
present in Saskatchewan, Canada
fecal sample (1 of 25, i.e., 4% of the samples) (Table 1). Since
No. positive for Bacteroidales
the abundance in the pigeon sample was low (3.3 log
copies g⫺1 feces) and detection occurred late in the qPCR
(with a threshold cycle [C
] value of 37.1), it is unlikely that
this false amplification would negatively impact the use of the
Individual human feces
assay as a tool for detection of Canada goose-specific fecal
pollution in environmental samples. In comparison, the nested
f5 assay described by Lu and colleagues (15)
demonstrated non-host-specific DNA amplification with fecal
DNA samples from several animals, including samples from
humans, pigeons, gulls, and agriculturally relevant pigs and
chickens (Table 1).
assays showed limits of
quantification (less than 10 copies of target DNA per reaction)similar to those of other host-specific Bacteroidales
qPCR assays (14, 16, 18). The sensitivities of the CGOF1-Bac
assays were 57% (with 58 of 101 samples
testing positive) and 50% (with 51 of 101 samples testing pos-itive) for Canada goose feces, respectively (Table 1). A similar
sensitivity of 58% (with 59 of 101 samples testing positive) was
obtained using the CG-Prev
f5 PCR assay. The combined use
of the three assays increased the detection level to 72% (73 of101) (Fig. 2). Importantly, all markers were detected within
groups of Canada goose feces collected each month from May
The 6 goose samples that tested negative for the All-Bac
marker also tested
to September, indicating relative temporal stability of the
negative for the three goose markers.
markers. The CG-Prev
f5 PCR assay is an end point assay, andtherefore the abundance of the gene marker in Canada goosefecal samples could not be determined. However, development
and Fig. 1). The high prevalence and abundance of Bacteroi-
of the CGOF1-Bac
qPCR approach allowed
in Canada goose feces suggested that detecting members
for the quantification of the host-specific CGOF1-Bac
of this order could be useful in identifying fecal contamination
markers. In the feces of some individual Canada
associated with Canada goose populations.
The composition of the Bacteroidales
community in Canada
goose feces (n
⫽ 15) was found to be relatively diverse since 52OTUs (with a cutoff of 98% similarity) were identified among211 nonchimeric 16S rRNA gene sequences. Phylogeneticanalysis of the 52 OTUs (labeled CGOF1 to CGOF52) re-vealed that 43 (representing 84% of the 16S rRNA gene se-quences) were Bacteroides
like and that 9 (representing 16% ofthe 16S rRNA gene sequences) were likely to be members ofthe Prevotella-
specific cluster (see Fig. S2 in the supplementalmaterial). Similarly, Jeter et al. (11) reported that 75.7% of theBacteroidales
16S rRNA clone library sequences generatedfrom goose fecal samples were Bacteroides
like. The majority ofthe Bacteroides
- and Prevotella
-like OTUs were dispersedamong a wide range of previously characterized sequencesfrom various hosts and did not occur in distinct clusters suit-able for the design of Canada goose-associated real-time quan-titative PCR (qPCR) assays (see Fig. S2 in the supplementalmaterial). However, two single Bacteroides
-like OTU se-quences (CGOF1 and CGOF2) contained putative goose-spe-cific DNA regions that were identified by in silico
analysis(using BLASTN, the probe match program of the RibosomalDatabase Project, and multiple alignment). The primers andprobe for the CGOF1-Bac
assays were de-signed with no mismatches to the clones CGOF1 and CGOF2,
FIG. 1. Concentrations of the Bacteroidales
) genetic markers in feces from various individual
assay demonstrated no cross-amplification
FREMAUX ET AL.
APPL. ENVIRON. MICROBIOL.
FIG. 2. Venn diagram for Canada goose fecal samples testing pos-
itive with the CGOF1-Bac
, and/or CG-Prev
f5 PCR as-
say. The number outside the circles indicates the number of Canada
goose fecal samples for which none of the markers were detected.
geese, the concentrations of CGOF1-Bac
were high, reaching levels up to 8.8 and 7.9 log
respectively (Fig. 1).
The potential of the Canada goose-specific Bacteroides
qPCR assays to detect Canada goose fecal pollution in an
environmental context was tested using water samples col-
lected weekly during September to November 2009 from 8
shoreline sampling sites at Wascana Lake (see File S1 and Fig.
S1 in the supplemental material). Wascana Lake is an urban
lake, located in the center of Regina, that is routinely fre-
quented by Canada geese. In brief, a single water sample of
approximately 1 liter was taken from the surface water at
each sampling site. Each water sample was analyzed for
enumeration using the Colilert-18/Quanti-
Tray detection system (IDEXX Laboratories, Westbrook,
ME) (8) and subjected to DNA extraction (with a PowerSoil
DNA extraction kit [Mo Bio Inc., Carlsbad, CA]) for the
detection of Bacteroidales
16S rRNA genetic markers using
) qPCR assay (14),
the two Canada goose-specific (CGOF1-Bac
) qPCR assays developed in this study, and the human-
specific (BacH) qPCR assay (17). All real-time and conven-
tional PCR procedures as well as subsequent data analysis are
described in the supplemental material and methods. The E.
quantification data demonstrated that Was-
cana Lake was regularly subjected to some form of fecal pol-
lution (Table 2). The All-Bac
genetic marker was consistently
detected in high concentrations (6 to 7 log
copies 100 ml⫺1)
in all the water samples, while E. coli
according to the sampling dates and sites, ranging from 0 to a
most probable number (MPN) of more than 2,000 100 ml⫺1.
High concentrations of E. coli
were consistently observed when
near-shore water experienced strong wave action under windy
conditions or when dense communities of birds were present at
a given site and time point.
The frequent detection of the genetic markers CGOF1-Bac
(in 65 of 75 water samples [87%]), CGOF2-Bac
(in 55 of 75
samples [73%]), and CG-Prev
f5 (in 60 of 75 samples [79%])
and the infrequent detection of the human-specific Bacteroi-
16S rRNA gene marker BacH (17) (in 5 of 75 water
samples [7%[) confirmed that Canada geese significantly con-
tributed to the fecal pollution in Wascana Lake during the
ESTIMATION OF CANADA GOOSE FECAL POLLUTION IN WATER
sampling period. Highest mean concentrations of both
birds regulatory report, no. 25. Population status of migratory game birds in
markers were obtained at the
Canada: November 2008. Canadian Wildlife Service, Gatineau, Quebec,Canada.
sampling sites W3 (3.8 and 3.9 log
copies 100 ml⫺1) and W4
5. Cole, J. R., Q. Wang, E. Cardenas, J. Fish, B. Chai, R. J. Farris, A. S.
copies 100 ml⫺1 for both), which are heavily fre-
Kulam-Syed-Mohideen, D. M. McGarrell, T. Marsh, G. M. Garrity, and
J. M. Tiedje.
2009. The Ribosomal Database Project: improved alignments
quented by Canada geese (Table 2), further confirming their
and new tools for rRNA analysis. Nucleic Acids Res. 37:
significant contribution to fecal pollution at these particular
6. Fallacara, D. M., C. M. Monahan, T. Y. Morishita, and R. F. Wack.
sites. It is worth noting that concentrations of the CGOF1-Bac
Fecal shedding and antimicrobial susceptibility of selected bacterial patho-gens and a survey of intestinal parasites in free-living waterfowl. Avian Dis.
markers in water samples displayed a signif-
icant positive relationship with each other (correlation coeffi-
7. Feare, C. J., M. F. Sanders, R. Blasco, and J. D. Bishop.
1999. Canada goose
cient ⫽ 0.87; P
⬍ 0.0001), supporting the accuracy of both
) droppings as a potential source of pathogenic bacteria.
J. R. Soc. Promot. Health 119:
assays for identifying Canada goose-associated fecal pollution
8. Fremaux, B., T. Boa, A. Chaykowski, S. Kasichayanula, J. Gritzfeld, L.
Braul, and C. Yost.
2009. Assessment of the microbial quality of irrigation
water in a prairie watershed. J. Appl. Microbiol. 106:
In conclusion, the CGOF1-Bac
9. Graczyk, T. K., R. Fayer, J. M. Trout, E. J. Lewis, C. A. Farley, I. Sulaiman,
says developed in this study are efficient tools for estimating
and A. A. Lal.
sp. cysts and infectious Cryptosporidium parvum
freshwater fecal inputs from Canada goose populations. Pre-
oocysts in the feces of migratory Canada geese (Branta canadensis
Environ. Microbiol. 64:
liminary results obtained during the course of the present study
10. Hudson, C. R., C. Quist, M. D. Lee, K. Keyes, S. V. Dodson, C. Morales, S.
also confirmed that Canada geese can serve as reservoirs of
Sanchez, D. G. White, and J. J. Maurer.
2000. Genetic relatedness of Sal-
species (see Fig. S3 in the sup-
isolates from nondomestic birds in southeastern United States.
J. Clin. Microbiol. 38:
plemental material). Therefore, future work will investigate
11. Jeter, S. N., C. M. McDermott, P. A. Bower, J. L. Kinzelman, M. J. Bootsma,
the cooccurence of these enteric pathogens with the Canada
G. W. Goetz, and S. L. McLellan.
diversity in ring-billed
goose fecal markers in the environment.
gulls (Laurus delawarensis
) residing at Lake Michigan beaches. Appl. Envi-
ron. Microbiol. 75:
Nucleotide sequence accession numbers.
12. Kassa, H., B. J. Harrington, and M. S. Bisesi.
2004. Cryptosporidiosis: a
16S rRNA gene sequences have been submitted to GenBank
brief literature review and update regarding Cryptosporidium
in feces of
Canada geese (Branta canadensis
). J. Environ. Health 66:
under accession numbers from GU222166 (clone CGOF1) to
13. Kullas, H., M. Coles, J. Rhyan, and L. Clark.
2002. Prevalence of Escherichia
GU222217 (clone CGOF52).
serogroups and human virulence factors in faeces of urban Canada geese
). Int. J. Environ. Health Res. 12:
14. Layton, A., L. McKay, D. Williams, V. Garrett, R. Gentry, and G. Sayler.
We acknowledge the generous support of Saskatchewan Agriculture
2006. Development of Bacteroides
16S rRNA gene TaqMan-based real-time
through the Agriculture Development Fund. This work was also sup-
PCR assays for estimation of total, human, and bovine fecal pollution in
ported by funding through the Canada Research Chairs program to
water. Appl. Environ. Microbiol. 72:
15. Lu, J., J. W. Santo Domingo, S. Hill, and T. A. Edge.
We also acknowledge Chris Somers and Arjunn Sachdev, who have
diversity and host-specific sequences of Canada goose feces. Appl. Environ.
assisted in the collection of fecal samples, and thank Jared Clarke of
16. Mieszkin, S., J. P. Furet, G. Corthier, and M. Gourmelon.
the Wascana Centre Authority for supplying the aerial photo displayed
of pig fecal contamination in a river catchment by real-time PCR using two
in the supplemental material.
16S rRNA genetic markers. Appl. Environ. Micro-
17. Reischer, G. H., D. C. Kasper, R. Steinborn, A. H. Farnleitner, and R. L.
1. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman.
2007. A quantitative real-time PCR assay for the highly sensitive and
Basic local alignment search tool. J. Mol. Biol. 215:
specific detection of human faecal influence in spring water from a large
2. Bernhard, A. E., and K. G. Field.
2000. A PCR assay to discriminate human
alpine catchment area. Lett. Appl. Microbiol. 44:
and ruminant feces on the basis of host differences in Bacteroides
18. Reischer, G. H., D. C. Kasper, R. Steinborn, R. L. Mach, and A. H. Farn-
genes encoding 16S rRNA. Appl. Environ. Microbiol. 66:
2006. Quantitative PCR method for sensitive detection of ruminant
3. Bernhard, A. E., and K. G. Field.
2000. Identification of nonpoint sources of
fecal pollution in freshwater and evaluation of this method in alpine karstic
fecal pollution in coastal waters by using host-specific 16S ribosomal DNA
regions. Appl. Environ. Microbiol. 72:
genetic markers from fecal anaerobes. Appl. Environ. Microbiol. 66:
19. Wahlstrom, H., E. Tysen, E. Olsson Engvall, B. Brandstrom, E. Eriksson, T.
Morner, and I. Vagsholm.
2003. Survey of Campylobacter
4. Canadian Wildlife Service Waterfowl Committee.
2008. CWS migratory
O157 and Salmonella
species in Swedish wildlife. Vet. Rec. 153:
20070195- Final Report Additional Info.pdf
Tetracycline Treatment Retards the Onset and Slows theProgression of Diabetes in Human Amylin/Islet AmyloidPolypeptide Transgenic MiceJacqueline F. Aitken,1,2 Kerry M. Loomes,1,2 David W. Scott,1,3 Shivanand Reddy,1Anthony R.J. Phillips,1,2,4 Gordana Prijic,1 Chathurini Fernando,1 Shaoping Zhang,1,2Ric Broadhurst,5 Phil L'Huillier,5 and Garth J.S. Cooper1,2,3,6
Lopez et al. BMC Medicine 2014, 12:200http://www.biomedcentral.com/1741-7015/12/200 Medicine for Global Health Remembering the forgotten non-communicablediseases Alan D Lopez1*, Thomas N Williams2,3, Adeera Levin4, Marcello Tonelli5, Jasvinder A Singh6,7,8, Peter GJ Burney9,Jürgen Rehm10,11,12,13,14, Nora D Volkow15, George Koob16 and Cleusa P Ferri17,18