Impact of land-use and fecal contamination on Escherichia populations in environmental samples

Abstract

Understanding the composition of complex Escherichia coli populations from the environment is necessary for identifying strategies to reduce the impacts of fecal contamination and protect public health. Metabarcoding targeting the hypervariable gene gnd was used to reveal the complex population diversity of E. coli and phenotypically indistinct Escherichia species in water, soil, sediment, aquatic biofilm, and fecal samples from native forest and pastoral sites. The resulting amplicons were cross-referenced against a database containing over 700 different partial gnd sequences from E. coli/non-E. coli Escherichia species. Alpha and beta measures of diversity of Escherichia populations were lowest in feces, soil and sediment compared to water and aquatic biofilm samples. Escherichia populations recovered from extensive freshwater catchments dominated by sheep, beef and dairy farming were extremely diverse but well-separated from a wetland dairy site. In contrast, Escherichia populations from the low-impact native forest site with fewer fecal sources were less diverse. Metabarcoding distinguished E. coli populations important to fecal contamination monitoring from non-E. coli Escherichia environmental populations. These data represent in-depth analysis and geographic stability of Escherichia populations from environmental samples with extensive heterogeneity, and reveal links with diverse fecal sources, land-use and the overall burden of fecal contamination at sample sites.

Keywords: Fecal indicator bacteria; Freshwater; Land-use; Microbial community profiling; Source tracking; Zoonotic pathogens.

Fig. 1

Sample site catchments included in study. Sites 1 to 4 are part of the broader Manawatū River catchment in the Tararua region of New Zealand. Site 5 is associated with a constructed wetland (circled in yellow), which receives tile drainage from a small catchment area (main area outlined in red), constructed on a dairy farm in the Waikato region. Sample sites indicated as white circles. Site 3 and Site 4 sampling locations were separated by less than 100 m so overlap. Background aerial imagery for Site 5 sourced from Toitū Te Whenua LINZ CC BY 4.0 Imagery Basemap contributors.

Fig. 2

Agglomerated reads of the top 20 gnd sequence types (gSTs) from the five sample types across individual sites. Total numbers of reads for the top 20 gSTs across 168 libraries (> 100 reads) were examined across sample types for each site. Numbers in blue represents the number of separate sample type libraries per site for which barcoded gnd amplicons were obtained. gST536, gST537 and gST548 are Escherichia marmotae; other gSTs are E. coli associated. Sites 1, 2 and 5 were visited on six occasions and Sites 3 and 4 on five occasions. No sediment samples were obtained for Site 2.

Fig. 3

Relationship between E. coli concentration and total number of gSTs representing E. coli and non-E. coli Escherichia species. A log-linear regression model was used to compare generic ‘E. coli’, enumerated (MPN per 100 mL water) using Colilert-18^®/Quanti-Tray/2000^®, and E. coli and non-E. coli Escherichia gnd sequence types (gSTs) detected using metabarcoding sequencing methods from water samples obtained at catchment sites dominated by native forest (Site 1), mixed sheep, beef and dairy farming (Site 2, 3 and 4), and dairy farming (Site 5). Community profiling was undertaken with gnd amplicon metabarcoding of water samples to identify total number of gSTs matching E. coli, and Escherichia species (E. marmotae, E. ruysiae and E. whittamii) within a reference database (gndDb), per sample.

Fig. 4

Alpha diversity analysis of data of Escherichia coli, and Escherichia species (E. marmotae, E. ruysiae and E. whittamii). Chao1 and Shannon index diversity analysis of data to examine variation of all gnd sequence types (gSTs) (n = 426) matching Escherichia coli, and Escherichia species (E. marmotae, E. ruysiae and E. whittamii) within a reference database (gndDb). Box and whisker plot of Chao1 (richness) and Shannon diversity indices of community profiles of 168 environmental samples examined using gnd amplicon metabarcoding. Amplicon sequencing variants were identified using DADA2 and compared to gnd sequence types within gndDb, a database containing over 700 hypervariable partial gnd sequences from freshwater Escherichia species. The boxes show median values and span lower to upper quartiles; the whiskers show the highest and lowest values within 1.5 times the interquartile range. A Student’s t-test of variance was used to compare (A) Chao1 and (B) Shannon diversity indices from Site 1 with each of the other sites, and (C) Chao1 and (D) Shannon diversity indices from water samples with other environmental samples; **, p < 0.01, ***, p < 0.001; ****, p < 0.0001; ns, not significant, p > 0.05. A one-way ANOVA was used to compare gST (n = 426) read counts from all sites or sample types.

Fig. 5

Alpha diversity analysis of non-E. coli Escherichia species. Chao1 and Shannon index diversity analysis of data to examine variation of all 22 non-E. coli Escherichia species amplicon sequencing variants across environmental samples and sample sites. Box and whisker plot of Chao1 (richness) and Shannon diversity indices of only E. marmotae, E. ruysiae and E. whittamii gnd sequence types (gSTs) identified within community profiles of 168 environmental samples examined using gnd amplicon metabarcoding. Amplicon sequencing variants were identified using DADA2 and compared to E. marmotae, E. ruysiae and E. whittamii gnd sequence types within gndDb, a database containing over 700 hypervariable partial gnd sequences. The boxes show median values and span lower to upper quartiles; the whiskers show the highest and lowest values within 1.5 times the interquartile range. A Student’s t-test of variance was used to compare (A) Chao 1 and (B) Shannon diversity indices of E. marmotae, E. ruysiae and E. whittamii gSTs from Site 1 with each of the other sites, and (C) Chao 1 and (D) Shannon diversity indices of E. marmotae and E. ruysiae gSTs from water samples with other environmental samples; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant, p > 0.05. A one-way ANOVA was used to compare all ASV read counts (n = 22) from all sites or sample types.

Fig. 6

Beta diversity of all Escherichia coli and non-E. coli Escherichia species gnd sequence types (gSTs). Principal component analysis (PCA) to demonstrate the between sample variation and spatial distribution of gSTs (n = 426) matching E. coli, and Escherichia species (E. marmotae, E. ruysiae and E. whittamii) within a reference database (gndDb) obtained from environmental samples recovered from five separate sites.