Distribution and Diversity of Escherichia coli Populations in the South Nation River Drainage Basin,Eastern Ontario,Canada

We investigated the prevalence and diversity of Escherichia coli strains isolated from surface waters from multiple watersheds within the South Nation River basin in eastern Ontario, Canada. The basin is composed of mixed but primarily agricultural land uses. From March 2004 to November 2007, a total of 2,004 surface water samples were collected from 24 sampling sites. E. coli densities ranged from undetectable to 1.64 × 10⁵ CFU 100 ml⁻¹ and were correlated with stream order and proximity to livestock production systems. The diversity of 21,307 E. coli isolates was characterized using repetitive extragenic palindromic PCR (rep-PCR), allowing for the identification of as many as 7,325 distinct genotypes, without capturing all of the diversity. The community was temporally and spatially dominated by a few dominant genotypes (clusters of more than 500 isolates) and several genotypes of intermediary abundance (clustering between 10 and 499 isolates). Simpson diversity indices, assessed on a normalized number of isolates per sample, ranged from 0.050 to 0.668. Simpson indices could be statistically discriminated on the basis of year and stream order, but land use, discharge, weather, and water physical-chemical properties were not statistically important discriminators. The detection of Campylobacter species was associated with statistically lower Simpson indices (greater diversity; P < 0.05). Waterborne E. coli isolates from genotypes of dominant and intermediary abundance were clustered with isolates obtained from fecal samples collected in the study area over the same period, and 90% of the isolates tested proved to share genotypes with fecal isolates. Overall, our data indicated that the densities and distribution of E. coli in these mixed-use watersheds were linked to stream order and livestock-based land uses. Waterborne E. coli populations that were distinct from fecal isolates were detected and, on this basis, were possibly naturalized E. coli strains.Escherichia coli is ubiquitously distributed in fecal material from humans and warm-blooded animals (38). The detection of E. coli in water is an implicit indicator of recent fecal contamination and therefore of the risk of cooccurrence of enteric pathogens that can cause illness in susceptible populations (62). Many jurisdictions evaluate and mandate compliance with drinking and recreational water quality standards on the basis of the presence and abundance of E. coli (14, 44). For example, Canadian recreational water quality standards stipulate that E. coli densities in excess of a geometric mean of 200 CFU per 100 ml indicate that the water is unsuitable for swimming and bathing (23).In a background of increasing occurrence of microbial contamination of surface water, a variety of methods for elucidating the sources of fecal contamination have been developed, and these microbial source tracking (MST) methods are recommended components of fecal pollution abatement strategies (16, 57). So-called library-dependent MST methods compare environmental isolates to collections of isolates obtained from likely sources of fecal pollution in the area of investigation. The host source is distinguished on the basis of the similarity of environmental isolates to reference fecal isolates. Comparison can be undertaken on the basis of genomic fingerprinting methods, including repetitive extragenic palindromic PCR (rep-PCR), ribotyping, or pulsed-field gel electrophoresis (PFGE) (13, 17, 31, 54, 57). A variety of studies using these methods have revealed enormous diversity in the fecal and environmental E. coli populations. For example, 461 distinct PFGE genotypes and 175 distinct enterobacterial repetitive intergenic consensus (ERIC)-PCR genotypes were detected in a collection of 555 E. coli strains isolated from river water in Texas (10). As many as 291 and 94 rep-PCR genotypes were distinguished in collections of 643 river isolates and 353 beach water E. coli isolates, respectively (43). Significant diversity was also revealed using multilocus enzyme electrophoresis (MLEE) and multilocus sequence typing (MLST) on 185 E. coli isolates from freshwater beaches, where an average of 40 alleles per locus were detected (59). Almost 60% of 657 E. coli isolates in a fecal reference collection had unique (i.e., detected in only one individual) fingerprints determined by rep-PCR (32). Extensive diversity of E. coli was also observed in soils in temperate climates, where the growth and persistence of “naturalized” populations without any known fecal input have been found (7, 28, 30). Naturalized populations have been dominated by the B1 phylogroup and may have adapted in ways that enhance their survival in temperate secondary habitats (59). The temporal and spatial diversity of E. coli may not be a significant factor in coarse-source (e.g., human versus animal) classification of E. coli by means of ribotyping procedures (48). Ultimately, the characterization and understanding of the diversity of populations of selected microorganisms in surface watercourses affected by multiple sources of fecal pollution (as in agricultural watershed settings, for example) may be more critical for assessing the specific impacts of contamination-mitigating measures than previously thought. For instance, restricting the access of cattle on pasture to adjacent water by implementing vegetative buffering along watercourses creates habitat for varied wildlife, which then contribute to fecal pollution. In this context, the diversity in populations of indicator bacteria could be useful for better understanding how changes in landscape use influence fecal source inputs.As part of a research program evaluating the impact of agriculture on water quality and the efficacy of better agricultural management practices to mitigate agricultural pollution, we have conducted a multiyear study of the microbiological water quality for a suite of different-sized watersheds in the South Nation River basin in eastern Ontario, Canada (41, 46, 61). Land use in this river basin is mixed, consisting primarily of agricultural activities, light urban development, and interspersed wildlife habitat. Surface water systems in the study region differ widely in their contributing areas and therefore in their discharges (61).In the work undertaken here, we sought to determine the spatial and seasonal variability in the density and the structure of populations of E. coli in surface waters within the South Nation River basin. The specific objectives of the study were (i) to characterize the seasonal distribution and abundance of E. coli in different watershed settings within the river basin, (ii) to evaluate the spatial distribution of E. coli densities and diversity with respect to upstream land use activities, (iii) to use rep-PCR to elucidate the dominant E. coli genotypes and the diversity of E. coli populations and to explore linkages to pathogen presence, season, and environmental and land use variables, and (iv) using rep-PCR, to evaluate the concordance between waterborne isolates and fecal isolates obtained from within the study area. The study is distinguished by an intensive 4-year sampling of numerous (n = 24) sites that differed in their stream order and proximal land use activity; the number of E. coli isolates (≈21,000) included in the analysis; and the use of two distinct rep-PCR fingerprinting methods (ERIC and BOXA1R) to characterize the isolates. Furthermore, we used classification and Regression Tree (CART) analysis to evaluate relationships between the abundance and diversity of E. coli in water samples and environmental and land use variables.