Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2011 Aug 11.
Published in final edited form as: J Appl Microbiol. 2010 Jul 9;109(5):1599–1608. doi: 10.1111/j.1365-2672.2010.04788.x

Escherichia coli and enterococci are sensitive and reliable indicators for human, livestock and wildlife faecal pollution in alpine mountainous water resources

AH Farnleitner 1,*, G Ryzinska-Paier 1, GH Reischer 1, MM Burtscher 1, S Knetsch 2, AKT Kirschner 2, T Dirnböck 3, G Kuschnig 4, RL Mach 1, R Sommer 2
PMCID: PMC3154642  EMSID: UKMS35832  PMID: 20629798

Abstract

Aims

This study evaluated the applicability of standard faecal indicator bacteria (SFIB) for alpine mountainous water resources monitoring.

Methods and Results

Escherichia coli, enterococci (ENTC) and Clostridium perfringens were investigated by standard or frequently applied methods in a broad range of animal and human faecal sources in a large alpine mountainous area. C. perfringens occurred only in human, livestock and carnivorous source groups in relevant average concentrations (log 4.7-7.0 CFU per g) but not in herbivorous wildlife sources. E. coli proved to be distributed in all faecal source groups with remarkably balanced average concentrations (log 7.0 - 8.4 CFU per g). Except for single faecal samples from the cattle source group, prevalence rates for ENTC source groups were generally > 90% with average concentrations of log 5.3 - 7.7 CFU per g. In order to test the faecal indication capacity in the environment, faecal prevalence data were comparatively analysed with results from the concurrently performed multi-parametric microbial source tracking effort on karst spring water quality from the investigated alpine mountainous catchment (Reischer et al. 2008; Environ. Microbiol. 10:2598-2608).

Conclusion

E. coli and enterococci are reliable faecal indicators for alpine mountainous water resources monitoring, although E. coli is the more sensitive one. C. perfringens did not prove an indicator of general faecal pollution but is suggested a conservative microbial source tracking marker for anthropogenic faecal influence.

Significance and Impact of Study

Applicability of SFIB is currently hotly debated. This is the first study providing comprehensive information on the applicability of SFIB at alpine mountainous habitats.

Keywords: Escherichia coli, Clostridium perfringens, enterococci, general faecal pollution indicators, microbial source tracking marker, water quality monitoring

Introduction

As much as 20% to 25% of the global population depends largely or entirely on karst water with a significant part originating from mountainous area (Ford and Williams, 2007). The Austrian water supply is covered by approximately 50% from such type of alpine mountainous karst water resource. Spring water issuing from aquifer areas with prevailing matrix-flow conditions reveals excellent raw water quality (Wilhartitz et al., 2009). However, during unfavourable conditions, e.g. caused by storm-flow runoff through large, temporarily activated karstic conduits (Ford and Williams, 2007), microbial water quality at the spring outlet can decrease rapidly (Stadler et al., 2008). With respect to the given hydro-geological situation efficient diagnostic tools for the detection and analysis of faecal pollution are needed to guide microbial system assessment, enable internal microbial water quality monitoring, or allow regulatory surveillance activities (WHO, 2004, Farnleitner et al., 2008).

Selective cultivation of standard faecal indicator bacteria (SFIB) is still the methodological and regulatory basis for faecal pollution analysis of water resources (WHO, 2004). Nonetheless, application of SFIB needs basic knowledge on their ecology in context to the primary (i.e. intestinal systems, faecal materials) and the secondary habitats (i.e. soil, sediment and water, cf. discussion for details). The use of SFIB for microbiological water quality monitoring has increasingly been debated during the last decade (Ishii and Sadowsky, 2008, Boehm et al., 2009). Unfortunately, the availability of comprehensive studies is very limited in order to judge the indication performance of SFIB on sound scientific criteria. Furthermore, performance characteristics have to be evaluated for well specified aquatic habitats and defined monitoring applications.

In relation to the primary intestinal habitat two fundamental requirements of SFIB are, i) presence in significant numbers in faeces in order to achieve reasonable detection limits, and ii) a balanced distribution of concentrations between animal and human faecal pollution sources to ensure un-discriminative detectability of total faecal pollution levels. Since the implementation of the faecal indicator concept more than 100 years ago, Escherichia coli - as the most reliable representative of the coliform group - intestinal enterococci (ENTC), and Clostridium perfringens have most frequently been used as SFIB (Slanetz and Bartley, 1957; Bonde, 1966; Geldreich et al., 1968; Geldreich and Kenner, 1969; Bisson and Cabelli, 1980; Leclerc et al., 2001). Much knowledge has been gained on the prevalence and abundance of E. coli, ENTC, and C. perfringens in human, sewage and animal faecal sources (e.g. Maki and Picard, 1965; Smith, 1965; Finegold et al., 1974; Yamagishi et al., 1976; Hausschild et al., 1979; Crane et al., 1983; Ferguson et al., 2009). However, a great deal of diverse cultivation techniques have been used for quantifying SFIB in faecal sources making comparisons with standard methods used for water quality monitoring difficult (e.g. ISO based protocols). Furthermore, investigations were mainly focused on human and livestock faecal sources neglecting abundant wildlife populations in rural areas irrespective of their potential relevance as zoonotic reservoirs (for review see Cotruvo et al., 2004). These gaps of knowledge appear very surprising given the importance of SFIB for quality monitoring and water resources management on a worldwide scale.

The aim of this research was i) to establish basic knowledge on the multiple prevalence and abundance of E. coli, ENTC and C. perfringens in human, livestock and wildlife faecal pollution sources in fresh to medium aged faecal material (≤ 3 weeks) as occurring in a large and representative alpine mountainous area in the Northern Calcareous Alps of Austria and, ii) to evaluate its faecal indication capacity in spring water of the corresponding catchment. For this purpose, a comprehensive set of faecal samples from dominating wildlife populations (chamois, red deer, roe deer), abundant livestock (cattle kept on summer pastures), human faecal sources (individual, sewage and cesspit samples from alpine huts and tourism activities) and less abundant populations (herbivorous and carnivorous wildlife) were investigated. Occurrence of SFIB was also concurrently determined in spring water from the respective alpine mountainous catchment together with a multi-parametric set of alternative microbial and source tracking parameters. Water quality analysis was paralleled by on-line hydrological spring water characterisation covering discharge at base and event flow situations. To our knowledge, this is the first comprehensive study on the quantitative occurrence of several SFIB including also wildlife pollution sources.

Materials and methods

Study area and file collection of specimens

The investigated alpine mountainous area is located in the Northern Calcareous Alps in Austria reaching altitudes up to approx. 2300 m with a total investigated area of approx. 100 km2. Vegetation comprises summer pastures, natural calcareous alpine swards with open krummholz and forests. The limestone karst aquifer spring 2 (LKAS2) catchment is part of the investigated area (catchment area approx. 70 km2) with a mean altitude of 1380 m above sea level and with LKAS2 accessible in the respective valley at 650 m. Since this area has been the focus of several detailed studies (geology, hydrogeology, vegetation, spring water quality) further background information can be found in the respective publications (references found in Farnleitner et al., 2005; Reischer et al., 2008). Two routes (A, B) were selected in order to cover all different alpine mountainous habitats available in the considered catchment. 14 and 15 field sampling tours were realised along sampling route A (total length 32 km) and B (total length 35 km), respectively, from 2004 to 2006. Sampling design was based on single faecal samples (SFS) and pooled faecal samples (PFS); PFS included 10 single samples of equal volume from closely located areas within a corresponding habitat type. About 5-8 g of faecal material was aseptically collected from faecal droplets or cow pats for SFS or PFS in sterile faecal sampling plastic containers (Greiner Bio-One, Kremsmuenster, Austria). A representative set of cesspit and sewage samples were also collected at each field trip from 10 different mountain huts scattered over the investigation area. In total, 162 SFS and 171 PFS from 1760 individual faecal specimens and 65 cesspit/sewage samples were collected. Additionally, 46 faecal samples from human individuals living in the eastern Austrian area (representing potential visiting tourists) were recovered. As the investigation was focusing on the occurrence of SFIB in fresh to medium aged faeces, still deformable animal faecal droplets, not yet completely dried, were collected. The age of collected faecal droplets was limited to a maximum of 3 weeks as the sampling frequency of field trips were performed every 2-3 weeks during summer time allowing to distinguish fresh/medium aged faeces from older deposits (> 3 weeks). Collected faecal droplets were transported under dark conditions and ambient temperatures in insulated sampling boxes to the laboratory, stored at 4°C and investigated within 72 h.

Faecal sample processing, SFIB enumeration and characterisation

Approximately 1 g of homogenised wet faecal sample, was weighed, suspended in 100 ml buffered peptone saline and diluted in ten-fold dilutions from 10−2 to 10−6, resulting in a detection limit ranging from 17 to 110 CFU per g of faeces (depending on the exact mass). Determination of E. coli was performed by surface plating technique using Trypton-Bile-X-Glucuronide (TBX) medium (Oxoid, UK) and incubation at 44±0.5°C for 44±4 h. Representative colonies (n = 98) were further isolated for genotypic characterisation. Genotypic identification was done by a species-specific PCR using the uidA gene region (uidA_PCR) as previously described (Bej et al., 1990). ENTC was performed by membrane filtration and subsequent dry heat incubation at 44±0.5°C for 44±4h using Slanetz-Bartley medium (SB, Oxoid, UK). Further characterisation was done by transferring membranes to Bile-Esculin agar (BEA, Oxoid, UK) for 2 h. Presumptive C. perfringens were enumerated by surface plating technique and incubation at 44±1°C for 21±3 h in anaerobic jars using Tryptose-Sulphite-Cycloserine medium (TSC, Scharlau, Spain). Species-specific PCR identification for representative isolates was based on the phospholipase C gene (plC_PCR) as described previously (Fach and Popoff, 1997).

Hydrological characterisation, water quality monitoring and statistics

A detailed description of the, i) selected sampling design including a three years basic monitoring program (2004-2006), ii) analysis of two late summer events (2005 and 2006), and iii) the hydrological and physicochemical LKAS2 characteristics was published in detail already by Reischer et al. (2008). In this study, which was conducted concurrently with the herein described work, a multi-parametric set of indicators was successfully applied at the LKAS2 for faecal source tracking. Basic data on E. coli, ENTC and C. perfringens are referenced from the work of Reischer et al. (2008) making a comparison on the occurrence of SFIB in faecal sources versus spring water possible. All statistical procedures were performed with SPSS version 17.0.

Results

Methodical performance characteristics

Genotypic identification was only performed for selected E. coli and C. perfringens colonies as presumptive ENTC were confirmed by the additional Bile-Esculin test (c.f. methods). No inhibition effects were discernable for the uidA_PCR and plC_PCR under the applied test conditions. A total amount of 96 out of 98 E. coli isolates (i.e. 98%), randomly chosen from TBX plates covering all faecal sources, were uidA_PCR positive. Because of the excellent specificity of the TBX method, recovered E. coli concentrations from faecal and sewage were handled as confirmed results for further analysis and data interpretation. In the case of C. perfringens, 41 out of 47 presumptive isolates (i.e. 87%) from human, cattle and carnivorous faecal sources could be identified as C. perfringens by plC_PCR. However, plC_PCR confirmed only 14 out of 34 presumptive isolates (i.e. 40%) from wildlife sources (chamois, red deer, roe deer) as C. perfringens. Presumptive C. perfringens concentrations from wildlife were divided by a factor of 2.5 (i.e. 40 positives among 100 tested) in order to convert to the anticipated level of confirmed results and to compensate for the false positives from TSC agar.

Median bulk densities of randomly selected faecal specimens did not differ between samples from human individuals (median 1.02 g per cm3, range 0.75 – 1.15 g per cm3, n = 34), red deer (median 1.04 g per cm3, range 0.82 – 1.17 g per cm3, n = 45), chamois (median 1.04 g per cm3, range 0.78 – 1.14 g per cm3, n = 72) and cattle (median 1.03 g per cm3, range 0.90 – 1.36 g per cm3, n = 70). No correlations between concentrations of SFIB versus bulk density of respective faecal samples were discernable, irrespective whether values below the detection thresholds were included or not (range of Spearman rho = |0.00 – 0.20|, n = 34 - 187).

Abundance of SFIB in faecal excrements

Except for two SFS (i.e. 0.5%), E. coli occurred ubiquitously in all investigated wildlife, livestock, and human individual faecal samples. In cesspits and sewage E. coli revealed somewhat lower prevalence rates of 88% and 93%, explainable by partial treatment with chlorine (Table 1).

Table 1.

Prevalence and abundance of Escherichia coli in faeces and sewage

Source Prevalence Abundance
n %
positiv		 log10
median
[CFU/g]5 		log10
mean1
[CFU/g]		 log10
percentile
[CFU/g]
5% 95%
Chamois SFS 52 98 6.2 7.2 3.6 8.1
PFS 62 100 6.8 7.8 4.5 8.5
Red Deer SFS 36 100 6.1 7.0 3.4 7.8
PFS 36 100 6.9 7.5 4.1 8.2
Cattle SFS 48 100 6.4 7.0 4.1 7.8
PFS 60 100 6.6 7.0 5.1 7.6
Human individual 46 100 6.4 8.0 4.3 8.9
cesspit 25 88 7.1 7.8 4.1 8.8
sewage 40 93 4.2 5.7 2.5 6.7
Roe Deer SFS 12 100 5.3 7.1 2.7 8.1
Carnivore2 SFS 12 100 8.3 8.4 6.5 8.9
Sheep/Goat SFS3 2 100 8.4 8.4 5.8 8.7
PFS4 3 100 7.1 7.6 6.2 8.1
Open in a new tab
1

arithmetic

2

samples were taken from 12 carnivores including martens, foxes, 1 cat and 1 dog

3

samples were taken from 2 sheeps

4

samples were taken from 2 sheeps and 1 goat

5

CFU/g: colony forming units per g

For chamois, red deer and cattle populations as well as human individuals average E. coli concentrations in positive samples showed remarkably low variations ranging from log10 7.0 CFU to log10 8.0 CFU per g faeces, irrespective of whether SFS or PFS were considered (Table 1). In contrast, a pronounced variation of the results from single determinations was evident. In this respect, ratios of the 95% to 5% percentiles varied up to log10 4.1 for PFS and up to log10 4.6 for SFS.

A high ENTC prevalence rate of 87% - 92% (SFS) and 95% - 100% (PFS) was discernable for the investigated wildlife populations (chamois, red-deer) as well as the human individual samples (96%, Table 2). ENTC were found in 92% and 100% of the cesspit and sewage samples. In cattle populations ENTC showed a decreased prevalence of 69% (SFS) and 91% (PFS), respectively. All carnivores and other herbivores shed ENTC.

Table 2.

Prevalence and abundance of enterococci (ENTC) in faeces and sewage

Source Prevalence Abundance
n %
positiv		 log10
median
[CFU/g]5 		log10
mean1
[CFU/g]		 log10
percentile
[CFU/g]
5% 95%
Chamois SFS 52 87 4.5 6.8 2.0 7.8
PFS 58 95 6.3 7.3 4.2 8.3
Red Deer SFS 36 92 4.9 6.0 2.5 7.1
PFS 35 100 5.3 6.7 3.4 7.8
Cattle SFS 48 69 4.1 5.5 2.2 6.3
PFS 55 91 4.2 5.3 2.0 6.3
Human individual 46 96 4.9 6.4 2.5 7.4
cesspit 25 92 6.9 7.7 4.3 8.7
sewage 40 100 4.6 5.3 3.2 5.8
Roe Deer SFS 12 92 4.0 6.4 2.6 7.2
Carnivore2 SFS 12 92 7.1 7.7 5.3 8.6
Sheep/Goat SFS3 2 100 7.3 7.3 7.3 7.3
PFS4 3 100 7.5 7.3 5.9 7.5
Open in a new tab
1

arithmetic

2

samples were taken from 12 carnivores including martens, foxes, 1 cat and 1 dog

3

samples were taken from 2 sheeps

4

samples were taken from 2 sheeps and 1 goat

5

CFU/g: colony forming units per g

For chamois, red deer and cattle populations as well as human individuals average ENTC concentrations in positive samples ranged from log10 5.3 CFU to log10 7.3 CFU per g faeces, with lowest average concentrations for cattle populations and highest for human chamois PFS (Table 2). For these sources the ratios of the 95% to 5% percentiles varied up to log10 4.3 (PFS) and log10 5.8 (SFS).

Prevalence of C. perfringens in faecal droplets of chamois and red deer populations was low to very low (4 - 28%). In contrast, C. perfringens revealed a markedly increased prevalence in faecal pats of ruminant livestock populations (63 - 88%) and for human individual (83%), cesspit (96%) and sewage (98%) pollution sources (Table 3).

Table 3.

Prevalence and abundance of C. perfringens in faeces and sewage

Source Prevalence Abundance
n %
positiv		 log10
median
[CFU/g]5 		log10
mean1
[CFU/g]		 log10
percentile
[CFU/g]
5% 95%
Chamois SFS 52 4 1.7 1.7 1.7 1.7
PFS 62 15 1.8 2.1 1.7 2.6
Red Deer SFS 36 28 2.2 2.8 1.5 3.3
PFS 36 25 2.3 3.4 1.6 4.2
Cattle SFS 48 63 3.8 4.9 1.8 5.8
PFS 60 88 4.1 4.7 2.5 5.4
Human individual 46 83 4.0 5.8 2.2 6.3
cesspit 25 96 4.7 5.5 2.7 6.2
sewage 40 98 3.8 4.2 1.6 4.9
Roe Deer SFS 12 25 3.1 5.2 1.7 5.7
Carnivore2 SFS 12 83 6.9 7.0 2.7 7.3
Sheep/Goat SFS3 2 100 2.6 2.6 2.6 2.7
PFS4 3 67 2.9 2.9 2.9 2.9
Open in a new tab
1

arithmetic

2

samples were taken from 12 carnivores including martens, foxes, 1 cat and 1 dog

3

samples were taken from 2 sheeps

4

samples were taken from 2 sheeps and 1 goat

5

CFU/g: colony forming units per g

In agreement with prevalence rates, average concentrations of C. perfringens for positive samples showed also low average values for chamois and red deer populations ranging from log10 1.7 - 2.8 CFU (SFS) and log10 2.1 - 3.4 CFU (PFS) per g faeces, whereas significantly increased average values (Mann Whitney U, p < 0.05) for cattle and human faecal sources (log10 4.7 – 5.8 CFU per g faeces) were obvious. For these sources, the ratios of the 95% to 5% percentiles varied up to log10 2.9 (PFS) and log10 4.1 (SFS). Interestingly, carnivorous faecal sources resulted in highest average C. perfringens concentrations of log10 7.0 CFU per g faeces (Table 3).

Low to very low correlations were found between the concentrations of E. coli, ENTC and C. perfringens in faecal excrements and cesspits (Fig 1a-c). Slightly higher correlations were only observed for E. coli vs. ENTC concentrations and ENTC vs. C. perfringens concentrations from sewage and cesspit samples, respectively (Fig. 1a, 1c).

Fig. 1.

Fig. 1

Open in a new tab

Correlation analysis of log E.coli versus log enterococci (ENTC) concentrations (a), log E.coli versus log Clostridium perfringens concentrations (b), and log enterococci versus log Clostridium perfringens concentrations (c) as found in wildlife, livestock and human pollution sources in the investigated catchment. Spearman correlation coefficients (ρ) were calculated. The diagonal shows the 1:1 ratio of the respective couples of indicators.

The median E. coli to ENTC ratios for faecal excreta including wildlife, livestock and individual human sources ranged from log 1.1 to log 2.3. Sewage and cesspit sources showed reduced median ratios of log −0.2 to log −0.3 (Table 4a). As a consequence of the low correlations between the different groups of SFIB, variation of single ratios was quite pronounced ranging over 6 orders of magnitude. Reasonable E. coli to C. perfringens ratios could only be calculated for human and cattle sources because of C. perfringens being absent in large parts of wildlife sources (Table 4b).

Table 4a.

E.coli to enterococci ratios

Source n log10
median		 log10
percentile
5% 95%
Chamois 98 1.1 −1.5 3.8
Red Deer 68 1.5 −1.6 4.3
Cattle 83 2.3 −0,1 4.2
Human individ. 44 1.7 −1.3 3.4
cesspit 22 −0.3 −2.3 2.3
sewage 37 −0.2 −1.9 2.3
Open in a new tab

n = replicate number

Table 4b.

E. coli to C. perfringens ratios

Source n log10
median		 log10
percentile
5% 95%
Chamois not possible to calculate
Red Deer not possible to calculate
Cattle 83 2.3 0.7 5.4
Human individ. 38 2.0 −4.6 5.3
cesspit 22 2.1 −1.1 4.2
sewage 36 1.0 −1.6 3.0
Open in a new tab

n = replicate number

Occurrence of SFIB in spring water

Results on the detection rate and on the abundance of SFIB as recovered from the investigation of the karstic spring LKAS2 remarkably reflected the general prevalence and abundance of SFIB as found in the faecal sources investigated in the catchment area (Table 5). For the seasonal spring monitoring approach, E. coli and ENTC could be detected in 66% of the investigated water samples with a trend towards higher E. coli concentrations. C. perfringens could only be detected in 24% of the investigated samples. Results from flood event investigations in the spring further supported this picture (Tab 5). E. coli and ENTC detection rates approached 100%, whereas C. perfringens was detected only in 20% of the investigated samples. The 95% percentiles amounted to 1490 CFU per litre, 391 CFU per litre, and 4 CFU per litre for E. coli, ENTC and C. perfringens, respectively (see discussion for details on the evaluation of the indication capacity at the spring water).

Table 5.

Detectable E. coli, enterococci (ENTC) and C. perfringens concentrations in spring water (LKAS2) during 2004-20061

E. coli ENTC C. perfringens
Prevalence Abundance Prevalence Abundance Prevalence Abundance
% [CFU/l] % [CFU/l] % [CFU/l]
n pos. median 5% 95% n pos. median 5% 95% n pos. median 5% 95%
Monitoring 53 66 3 0 420 53 66 2 0 141 51 24 0 0 3
Event 70 100 177 32 1490 70 100 50 6 391 70 20 0 0 4
Open in a new tab

Abbreviations: CFU/l: colony forming units per litre of spring water; range of 5% to 95% percentile; n = replicate number; pos.: number of positive detections.

1

Raw data set originating from Reischer et al. 2008 but analysed and presented in a different way to support the herein given work.

Discussion

Occurrence of SFIB in their primary intestinal habitat

To our knowledge, this is the first available report on a comparative quantitative investigation of E. coli and alternative SFIB in faecal sources including herbivorous wildlife populations. Furthermore standard- or frequently applied cultivation methods for bacterial water quality investigations were used in order to ensure comparability of results. E. coli was found to occur at remarkably balanced average concentrations regardless of whether human, livestock, or wildlife sources were investigated. The pronounced variability of E.coli concentrations between SFS has been reported for decades (e.g. Geldreich, 1978; Donnison et al., 2008). This variability can simply be explained by the natural microbial populations dynamics as observed in the intestinal system (Williams and Crabb, 1961). Since faecal droppings were up to three weeks old, die-off and re-growth of SFIB in the faecal material might also have shifted populations (Sinton et al., 2007). The observed variability of E. coli concentrations between SFS is not considered a practical limitation for water quality monitoring, since rigorous mixing processes during mobilisation and transfer in the catchment are expected. This effect was already anticipated in the application of the PFS approach leading to a significant decrease in variability as compared to SFS. Furthermore, a remarkably decreased variability in E. coli to ENTC ratios was also observed for spring water samples when compared to ratios in faecal samples again supporting the postulated mechanism (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

Correlation plot of log E.coli versus log enterococci (ENTC) concentrations as found in animal and human faecal pollution sources and in the respective spring water of the LKAS2. The diagonal shows the 1:1 ratio. Indicator concentrations from spring water are given in log CFU per 10 m3 of spring water.

It is currently not clear why approx. 1/3 of the investigated cow pats (SFS) did not carry ENTC. A detailed investigation on ENTC isolates from corresponding cattle faecal droplets at the LKAS2 catchment area during 2004 revealed that S. bovis, E. faecium, E. hirae and E. durans were the dominating isolates (Burtscher et al., 2006). Hence, a rapid decrease of ENTC in faecal droplets due to fast dying S. bovis populations (Geldreich and Kenner, 1969; Feachem, 1975) alone cannot explain the lower prevalence rate. Average ENTC concentrations were reasonably high and quite balanced for the rest of the investigated source groups (human, herbivorous wildlife, carnivorous wildlife). ENTC thus fulfils the basic criteria as a total faecal pollution indicator, although less excellently than E. coli.

Quite surprisingly, C. perfringens could only be detected in relevant concentrations in human, livestock and carnivorous faecal sources. For these source groups concentrations were in the range of previously reported values for C. perfringens (Bonde, 1963; Yamagishi et al., 1976; Cabelli, 1977) but below concentrations observed for E. coli and ENTC. Interestingly, highest concentrations of C. perfringens were found for carnivorous faecal sources. Analysis of the genome suggests C. perfringens to be an anaerobic fastidious pathogenic flesh eating organism (Shimizu et al., 2002). Occurrence of C. perfringens in faeces of carnivorous animals and also humans (feeding on protein rich meat, fish or cheese) in higher numbers may thus be a logic consequence of its ecology. The very low prevalence and almost neglectable abundance in herbivorous wildlife severely limits C. perfringens as an indicator of total faecal pollution. C. perfringens has been suggested by several researchers as an indicator of stream or surface water quality (Bonde, 1963; Bisson and Cabelli, 1980; Fujioka and Shizumura, 1985; Sartory, 1988; Sorensen et al., 1989). As those studies focused mainly on the detection of sewage influence our results do not necessarily contradict this previous work. Because C. perfringens was found in humans as well as in herbivorous livestock but not in herbivorous wildlife, we suggest C. perfringens as a conservative microbial source tracking marker for faecal pollution from anthropogenic sources. Preliminary data on the occurrence of C. perfringens in faeces of native roe deer populations versus populations held in fenced areas and fed on silage in the LKAS2 area further supported this hypothesis. Only faecal material from populations held in fenced areas contained significant C. perfringens concentrations (data not shown). It appears unlikely that the suggested indication capacity for anthropogenic faecal pollution might be impaired by faecal droplets from carnivorous wildlife sources because of the comparatively low abundance of such animals. However, further work needs to be done in other environments to corroborate this hypothesis.

Faecal indication capacity of SFIB at the alpine mountainous spring habitat

Observed loads of SFIB at karst springs are expected to be a result of faecal material input, microbial transfer efficiency and die-off, assuming that extra-intestinal reservoirs of SFIB do not exist or do not actively contribute to environmental fluxes of SFIB (e.g. soils with interfering autochthonous populations). Under such a situation, prevalence rates and concentration ratios in spring water should correspond to patterns of SFIB as found in the contributing faecal sources. Indeed, results on the detection rate and on the abundance of SFIB in the LKAS2 spring remarkably reflected the general prevalence and abundance of SFIB as found in the faecal sources. Furthermore, the observed median E. coli to ENTC ratios from LKAS2 (log 0.3 and 0.6 for monitoring and flood event data, respectively) were also in correspondence to the median E. coli to ENTC ratios from animal faecal sources (log 1.1 to 2.3). Due to a higher persistence of ENTC in the environment a decreasing E. coli to ENTC ratio with time of environmental exposure is expected (Feachem, 1975). In contrast, E.coli to ENTC ratios from LKAS2 were remarkably higher than ratios in waste water and cesspits (log −0.2 to −0.3). These results are in perfect agreement with the concurrently performed source tracking study identifying ruminants as the dominating faecal pollution sources in the LKAS2 catchment (Reischer et al., 2008). In this respect it should be highlighted, that E. coli to ENTC ratios do not contain any information on the origin of faecal contamination as their ratios in animal and human sources were not different at all (cf. Table 4a). However, supporting information for improved data interpretation - as done in our case - can be derived from such differential persistence kinetics of SFIB (Feachem, R., 1975).

Most importantly, the good faecal indication capability of E.coli and enterococci in the LKAS2 was strongly supported by a very tight coupling to the observed ruminant-specific faecal pollution component being clearly distinguishable from non-intestinal parameters (Reischer et al., 2008). Multi-parametric analysis (n = 12 microbial and physiochemical parameter) revealed the highest correlations between E. coli and ENTC concentrations for the monitoring setup (r = 0.84). For the flood events, the highest correlations were observed for the “inter-related” faecal parameter group including E. coli, ENTC and ruminant-specific Bacteroidetes marker BacR (r = 0.81 - 0.88). The tightly correlated faecal indicator group revealed lower correlations with other microbial and physicochemical indicators such as turbidity (r = 0.24 - 0.60) or aerobic spore formers (r = 0.17 - 0.67) - a potential indicator for soil influence - strongly pointing to the fact that the origin of E. coli and ENTC was from faecal pollution sources and not from soil habitats (cf. Reischer et al., 2008 for details). No signs of any indication bias for E.coli and enteroccoci could thus be found by the multi-parametric catchment and spring aquifer study.

The detection frequency of C. perfringens in the LKAS2 was very low and in line with its occurrence in the investigated faecal sources. Only during periods of strongest faecal pollution, such as during the peak phases of floods, a good correlation to other SFIB was discernible (data not shown).

Recent results from other geographical regions indicated that E. coli populations may become “naturalised” by sustaining autochthonous populations not only in tropical but also in some temperate environments (Ishii and Sadowsky, 2008). Almost no information exists for ENTC on this subject but the situation is likely to be similar (Fujioka et al., 1999). Due to the fastidious nature of C. perfringens (Shimizu et al., 2002), significant growth in the secondary environment seems not very likely (Konishi et al., 1981; Zaiss et al., 1983). The primary habitats are diseased or dead bodies of animals (i.e. flesh as a basic growth substrate), large amounts of C. perfringens cells and spores are probably produced in carcasses especially during periods of higher temperatures. However, E. coli could recently be detected in the top soil layer of alpine pasture grassland up to concentration levels of 104 MPN per gram soil (Texier et al., 2008). In contrast, E. coli could not be isolated in soil from areas not frequented by cattle (Texier et al., 2008). Own investigations in the LKAS2 catchment support these results showing faecal coliforms up to 2.6×103 CFU per g of soil from pasture areas as in contrast to non-pasture areas revealing not detectable- or very low concentrations (i.e. < 2 CFU per g of soil; Burtscher, 2002). Soil layers below areas with high faecal exposure, such as soil below cattle pastures, are thus likely to receive, store, or even support the existence of naturalised populations in low to moderate concentrations as compared to their primary, intestinal habitat. The occurrence of non-faecal derived E. coli or ENTC, especially at very low concentrations, can thus not be excluded in spring water sources. A relevant faecal indication bias is only expected when strong erosive run-off events from catchments with dominating pasturing activities would happen. Enteric pathogenic organisms - likely to be stored or sustained under intensively pastured areas - may be mobilised as well under such situations (Fremaux et al., 2010). However, in the course of the performed multi-parametric source tracking investigation no signs of any significant indication bias from E.coli or enterococci from non-intestinal compartments could be found. In contrast, these investigations proved the excellent applicability of these SFIB for water quality monitoring at alpine mountainous water resources.

In conclusion, this study impressively demonstrates that E. coli and enterococci occur ubiquitously in faeces from human and animal sources irrespectively whether livestock or wildlife is regarded. Although investigations were limited to wildlife as occurring in mountainous areas the result of the study is likely being applicable also to other herbivorous wildlife populations with comparable diet and physiology. C. perfringens did not proof to indicate general faecal pollution but it is suggested as a conservative source tracking marker for anthropogenic faecal pollution. C. perfringens occurred in all faecal sources associated with human activities. Further studies in other geographical areas are needed in order to verify this hypothesis. In any case, higher volumes of water have to be sampled in order to compensate for the lower abundance of C. perfringens in faeces as compared to E. coli or ENTC (cf. Table 4a/4b). Combined analysis of faecal prevalence data (i.e. primary habitat) with results from the concurrently performed multi-parametric microbial source tracking effort on the respective karst spring water quality (i.e. secondary habitat) revealed a very good capacity for E. coli and ENTC to indicate microbial faecal pollution. E.coli and enterococci thus represent very reliable and useful general faecal indicators for water resource monitoring at alpine mountainous areas. However, these results are probably very habitat specific and must not be transferred to other locations, such as to lower altitude regions, without verification. It is the opinion of the authors that the current debate on the applicability of SFIB can only be solved if monitoring applications are precisely defined and discussed in a distinguished way.

Acknowledgements

This study was funded and supported by the Vienna Waterworks. Special thanks go to members of the Vienna Waterworks and to members of the Agricultural Forestry Department of Vienna for the crucial support during field sampling.

References

  1. Bej AK, Steffan RJ, DiCesare J, Haff L, Atlas RM. Detection of coliform bacteria in water by polymerase chain reaction and gene probes. Applied Environmental Microbiology. 1990;56:307–314. doi: 10.1128/aem.56.2.307-314.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bisson JW, Cabelli VJ. Clostridium perfringens as a water pollution indicator. Journal WPCF. 1980;52:241–248. [PubMed] [Google Scholar]
  3. Bonde GJ. Bacterial indicators of water pollution: a study of quantitative estimation. Teknisk Forlag; Copenhagen: 1963. p. 422. [Google Scholar]
  4. Bonde GJ. Bacteriological methods for estimation of water pollution. Health Laboratory Science. 1966;3:124–128. [PubMed] [Google Scholar]
  5. Boehm AB, Ashbolt NJ, Colford JM, Dunbar LE, Fleming LE, Gold MA, Hansel JA, Hunter PR, Ichida AM, McGee CD, Soller JA, Weisberg SB. A sea change ahead for recreational water quality criteria. Journal of Water and Health. 2009;7:9–20. doi: 10.2166/wh.2009.122. [DOI] [PubMed] [Google Scholar]
  6. Burtscher MM. Vienna University of Technology; 2002. Differentiation and determination of bacterial fecal contamination by means of Escherichia coli in a representative karst catchment area; p. 92. Master Thesis. [Google Scholar]
  7. Burtscher MM, Köllner K, Sommer R, Keiblinger K, Farnleitner AH, Mach RL. Development of a highly reproducible amplified fragment length polymorphism (AFLP) method for typing of Enterococcus spp. and application on isolates from cattle faeces in an alpine watershed. Journal of Microbiological Methods. 2006;66:281–293. doi: 10.1016/j.mimet.2006.04.016. [DOI] [PubMed] [Google Scholar]
  8. Cabelli VJ. Clostridium perfringens as a water quality indicator. In: Hodleay AW, Dutka BJ, editors. Bacterial indicators / health hazards associated with water. American Society for Testing Materials; 1977. pp. 65–79. [Google Scholar]
  9. Cabelli VJ. Obligate anaerobic bacterial indicators. In: Berg G, editor. Indicators of viruses in water and food. Ann Abour Sciences; Ann Abour: 1978. pp. 171–200. [Google Scholar]
  10. Cotruvo JA, Dufour A, Rees G, Bartram J, Carr R, Cliver DO, et al., editors. Waterborn Zoonoses. IWA Publishing; London: 2004. p. 506. [Google Scholar]
  11. Crane SR, Moore JA, Grismer ME, Miner JR. Bacterial pollution from agricultural sources: a review. Transactions of the ASAE. 1983:858–872. [Google Scholar]
  12. Donnison A, Ross C, Clark D. Escherichia coli shedding by dairy cows. New Zealand Journal of Agriculture Research. 2008;51:273–278. [Google Scholar]
  13. Fach, Popoff Detection of enterotoxogenic Clostridium perfringens in food and faecal samples with a duplex PCR and the slide latex agglutination test. Applied Environmental Microbiology. 1997;63:4232–4236. doi: 10.1128/aem.63.11.4232-4236.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Farnleitner AH, Wilhartitz I, Ryzinska G, Kirschner AKT, Stadler H, Burtscher M, Hornek R, et al. Bacterial dynamics in spring water of two contrasting alpine karst aquifers indicate autochthounous microbial endokarst communities. Environmental Microbiology. 2005;7:1248–1259. doi: 10.1111/j.1462-2920.2005.00810.x. [DOI] [PubMed] [Google Scholar]
  15. Farnleitner AH, H. S, Reischer GH, Sommer R, Kirschner AKT, Burtscher MM, et al. Methods and strategies for alpine karstic water resource management: opening pollution microbiology’s “black box”. Presented at: World Water Conference and Exhibition 2008; Vienna: International Water Association; 2008. [Google Scholar]
  16. Feachem R. An improved role for faecal coliform to faecal streptococci ratios in the differentiation between human and non-human pollution sources. Water Research. 1975;9:689–690. [Google Scholar]
  17. Ferguson CM, Charles K, Deere D. Quantification of microbial sources in drinking-water catchments. Crititical Reviews in Environmental Science and Technology. 2009;39:1–40. [Google Scholar]
  18. Finegold MS, Attebery HR, Sutter VL. Effect of diet on human fecal flora: comparison of Japanese and American diets. The Amercian Journal of Clinical Nutritian. 1974;27:1456–1469. doi: 10.1093/ajcn/27.12.1456. [DOI] [PubMed] [Google Scholar]
  19. Ford DC, Williams P. Karst hydrogeology and geomorphology. Wiley; 2007. p. 620. [Google Scholar]
  20. Fremaux B, Prigent-Combaret C, Beutin L, Gleizal A, Trevisan D, Quetin P, Jocteur-Monrozier L, Rozand C. Survival and spread of Shiga toxin-producing Escherichia coli in alpine pasture grasslands. Journal of Applied Microbiology. 2010 doi: 10.1111/j.1365-2672.2009.04527.x. doi:10.1111/j.1365-2672.2009.04527.x. [DOI] [PubMed] [Google Scholar]
  21. Fujioka R, Sian-Denton C, Borja M, Castro J, Morphew K. Soil: the environmental souce of Escherichia coli and enterococci in Guams streams. Journal of Applied Microbiology. 1999;85:83S–89S. doi: 10.1111/j.1365-2672.1998.tb05286.x. [DOI] [PubMed] [Google Scholar]
  22. Fujioka RS, Shizumura LK. Clostridium perfringens a reliable indicator of stream water quality. Journal WPCF. 1985;57:986–992. [Google Scholar]
  23. Geldreich EE. Bacterial populations and indicator concepts in feces, sewage, stormwater and solid wastes. In: Berg G, editor. Indicators of virus in water and food. Ann Abor Science Publishers, Inc.; Ann Abor, MI: 1978. pp. 51–97. [Google Scholar]
  24. Geldreich EE, Kenner BA. Concepts of fecal streptococci in stream pollution. Journal WPCF. 1969;41:R336–R352. [PubMed] [Google Scholar]
  25. Geldreich EE, Best LC, Kenner BA, Van Donsel DJ. The bacteriological aspects of stormwater pollution. Journal WPCF. 1968;40:1861–1872. [PubMed] [Google Scholar]
  26. Hausschild AHW, Desmarchelier P, Gilbert RJ, Harmon SM, Vahlefeld R. ICMSF methods studies. XII. Comparative study for the enumeration of Clostridium perfringens in feces. Canadian Journal of Microbiology. 1979;25:953–963. doi: 10.1139/m79-146. [DOI] [PubMed] [Google Scholar]
  27. Hofer I. Universität Wien; Vienna: 2005. Die Alm- und Weidewirtschaft des Rax-, Schneeberg-, Schneealpe-, Hochschwabgebietes: Eine Darstellung der Geographischen Grundlagen; p. 120. Master Thesis. [Google Scholar]
  28. Ishii S, Sadowsky MJ. Escherichia coli in the environment: implications for water quality and human health. Microbes and Environments. 2008;23:101–108. doi: 10.1264/jsme2.23.101. [DOI] [PubMed] [Google Scholar]
  29. ISO . Water quality - detection and enumeration of Escherichia coli and coliform bacteria - Part 1: membrane filtration method (ISO 9308-1: 2000) International Organization of Standardization; Geneva, Switzerland: 2000. [Google Scholar]
  30. ISO . Water quality - detection and enumeration of intestinal Enterococci - Part 2: membrane filtration method (ISO 7899-2: 2000) International Organization of Standardization; Geneva, Switzerland: 2000. [Google Scholar]
  31. ISO . Water quality – detection and enumeration of Clostridium perfringens – Part 2: method by membrane filtration (ISO/CD 6461-2) International Organization of Standardization; Geneva, Switzerland: 2002. [Google Scholar]
  32. Konishi K, Yamagishi T, ishisaka S, Sakamoto K, Sakurai S. Incidence and survival of Clostridium perfringens in soil. Japanese Journal of Bacteriology. 1981;36:459–464. [PubMed] [Google Scholar]
  33. Leclerc H, Mossel DAA, Edberg SC, Struijk CB. Advances in the bacteriology of the coliform group: their suitability as markers of microbial water safety. Annual Review Microbiology. 2001;55:201–234. doi: 10.1146/annurev.micro.55.1.201. [DOI] [PubMed] [Google Scholar]
  34. Maki LR, Picard K. Normal intestinal flora of cattle fed high-roughage rations. Journal of Bacteriology. 1965;89:1244–1249. doi: 10.1128/jb.89.5.1244-1249.1965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pourcher AM, Devriese LA, Hernandez JF, Delattre JM. Enumeration by a miniaturized method of Escherichia coli, Streptococcus bovis and enterococci as indicators of the origin of faecal pollution. Journal of Applied Bacteriology. 1991;70:525–530. doi: 10.1111/j.1365-2672.1991.tb02752.x. [DOI] [PubMed] [Google Scholar]
  36. Reischer GH, Haider JM, Sommer R, Stadler H, Keiblinger KM, Hornek R, Mach RL, Farnleitner AH. Quantitative microbial faecal source tracking with sampling guided by hydrological catchment dynamics. Environmental Microbiology. 2008;10:2598–2608. doi: 10.1111/j.1462-2920.2008.01682.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sartory DP. Faecal clostridia and indicator bacteria levels in an eutrophic impoundment. Water SA. 1988;14:115–117. [Google Scholar]
  38. Shimizu T, Ohtani K, Hirakawa H, Ohshima K, Yamashita A, Shiba T, et al. Complete genome sequence of Clostridium perfringens, an anaerobic flesh-eater. PNAS. 2002;99:996–1001. doi: 10.1073/pnas.022493799. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sinton LW, Braithwaite RR, Hall CH, Mackenzie ML. Survival of indicator and pathogenic bacteria in bovine feces on pasture. Applied Environmental Microbiology. 2007;73:7917–7925. doi: 10.1128/AEM.01620-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Slanetz LW, Bartley CH. Numbers of enterococci in water, sewage, and feces determined by the membrane filtration technique with an improved medium. Journal of Bacteriology. 1957;57:591–595. doi: 10.1128/jb.74.5.591-595.1957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Smith H, Crabb WE. The faecal bacterial flora of animals and man: its development in the young. Journal Pathology Bacteriology. 1961;82:53–66. [Google Scholar]
  42. Sorensen DL, Eberl SG, Dicksa RA. Clostridium perfringens as a point source indicator in non-point polluted streams. Water Research. 1989;23:191–197. [Google Scholar]
  43. Stadler H, Skritek P, Sommer R, Mach RL, Zerobin W, Farnleitner AH. Microbiological monitoring and automated event sampling at karst springs using LEO-satellites. Water Science and Technology. 2008;58:899–909. doi: 10.2166/wst.2008.442. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Texier S, Prigent-Combaret C, Gourdon MH, Poirier MA, Faivre P, Dorioz JM, et al. Persistence of culturable Escherichia coli fecal contaminants in dairy alpine grassland soils. Journal Environmental Quality. 2008;37:2299–2310. doi: 10.2134/jeq2008.0028. [DOI] [PubMed] [Google Scholar]
  45. WHO . Guidlines for drinking-water quality. World Health Organisation; Geneva: 2004. [Google Scholar]
  46. Wilhartitz I, Kirschner AKT, Mach RLM, Stadler H, Herndl G, Farnleitner AH. Prokaryotic production in karstic alpine spring aquifers and their ecological implications. FEMS Microbiology Ecology. 2009;68:287–299. doi: 10.1111/j.1574-6941.2009.00679.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Williams H, Crabb WE. The fecal bacterial flora of animals and man: its development in the young. Journal of Pathology Bacteriology. 1961;82:53–66. [Google Scholar]
  48. Wilson K. Preparation of genomic DNA from bacteria. In: Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, editors. Current Protocols in Molecular Biology. Wiley; New York: 1994. pp. 2.4.1–2.4.2. [DOI] [PubMed] [Google Scholar]
  49. Yamagishi T, Serikawa T, Morita R, Nakamura S, Nishida S. Persistent high numbers of Clostridium perfringens in the intestines of Japanese aged adults. Japanese Journal Microbiology. 1976;20:397–403. doi: 10.1111/j.1348-0421.1976.tb01005.x. [DOI] [PubMed] [Google Scholar]
  50. Zaiss U, Unteregger B, Kaltwasser H. The distribution of the gas-gangrene causing C. perfingens in rural rivers. gwf-wasser/abwasser. 1983;124:536–540. (in german) [Google Scholar]

RESOURCES