Abstract
Aim
To determine the prevalence, antimicrobial resistance, and genomic characteristics of antimicrobial-resistant Escherichia coli isolated from recreational freshwater in the Greys-Hoback watershed, Wyoming, USA.
Methods and results
A total of 263 presumptive E. coli isolates were recovered using membrane filtration (EPA Method 1603), of which 218 (83%) were confirmed as E. coli by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Antimicrobial susceptibility testing by broth microdilution identified 20 (∼9%) isolates with clinically relevant resistance phenotypes, including six multidrug-resistant (MDR; ≥3 antimicrobial classes). Whole-genome sequencing of the resistant subset identified blaEC alleles in all genomes, along with additional resistance determinants, including sul2, tet, aph variants, and floR genes. Plasmid replicons were detected in most isolates, with IncF-type plasmids predominating and associated with MDR gene clusters. Contig-level analysis identified plasmid-associated MDR regions in selected isolates, although plasmid reconstruction was limited by short-read assemblies. Phylogenetic analyses revealed multiple sequence types and O: H serotypes, demonstrating substantial genetic diversity and suggesting that the resistant E. coli population originated from multiple sources rather than from clonal expansion. All sequenced isolates carried multiple virulence-associated genes, indicating co-occurrence of resistance and virulence determinants.
Conclusions
By integrating antimicrobial susceptibility testing, whole-genome sequencing, and plasmid reconstruction, we show that recreational freshwater in northwestern Wyoming harbors genetically diverse antimicrobial-resistant pathogenic E. coli, including MDR isolates associated with putative mobile resistance elements. These findings support the inclusion of recreational water systems in environmental AMR surveillance frameworks and highlight the value of integrating phenotypic and genomic approaches for isolate characterization.
Keywords: antimicrobial resistance, Escherichia coli, multidrug-resistant bacteria, plasmids, environmental/recreational water
Impact Statement.
Most antimicrobial resistance (AMR) surveillance focuses on clinical, agricultural, and wastewater settings, often overlooking environmental settings where people may interact with contaminated water. Recreational rivers are infrequently considered in AMR policy, yet they are places where humans, livestock, wildlife, and antibiotic residues intersect, making them crucial “zones” within the AMR network. The present study demonstrates a practical approach to transforming these sites into AMR indicators by combining standard water quality monitoring with whole-genome sequencing, providing an actionable framework for earlier detection of high-risk bacterial lineages, more informed recreation advisories, and targeted upstream interventions within a One Health framework.
Introduction
Antimicrobials have extended human life expectancy by at least 20 years by enabling effective treatment of bacterial infections. However, widespread antibiotic use has accelerated the emergence of antimicrobial resistance (AMR). The World Health Organization identifies AMR as one of the top ten global public health threats and projects that it will contribute substantially to global mortality and economic losses by 2050. These resistant bacteria are ubiquitous across diverse environments, including meat, wastewater, freshwater, and other environmental reservoirs (Antimicrobial Resistance Collaborators 2022, Boidya et al. 2025, Ghorbani Tajani et al. 2026, Kaur et al. 2026).
In addition to its impact on human health, AMR also imposes substantial economic challenges, increases healthcare costs, extends hospital stays, and reduces productivity (Kraemer et al. 2022). Antimicrobial resistance genes (ARGs) have been present in natural ecosystems for millennia as intrinsic defense mechanisms against naturally occurring antimicrobials present in plants and other environmental stressors (Perry et al. 2014). Their increasing detection in clinically important pathogens is particularly concerning, especially in regions with inadequate water, sanitation, and hygiene infrastructure. A 2015 survey in the Netherlands found that 17.1% of Extended-Spectrum Beta-Lactamases (ESBL)-producing E. coli from river water and wastewater were classified as pathogenic based on virulence profiling, and among these pathogenic isolates, ∼84% were resistant to up to three antibiotic classes (Franz et al. 2015). In 2019, a study from northern Colorado showed that, although total bacteria and E. coli abundances decreased through wastewater treatment, the relative proportion of ESBL-resistant E. coli increased by 1.7% in surface waters, with 70 isolates validated across 17 antibiotics and 185 AMR genes detected via whole-genome sequencing (WGS; Haberecht et al. 2019). Additionally, in Wyoming, Bacillus cereus strain IM8, isolated from an Upper Fremont Glacier ice core, displayed high cytotoxicity toward human intestinal epithelial Caco-2 cells and resistance to ampicillin (Chandross-Cohen et al. 2026). Together, these findings highlight the importance of monitoring environmental bacteria for AMR and virulence traits across diverse water sources and geographic regions.
Exposure through recreational activities, such as swimming, fishing, or agricultural irrigation, can lead to ingestion or contact, resulting in infections or colonization by resistant bacteria. In arid and semi-arid regions like the northwestern USA, AMR risks are amplified by agricultural activity and limited sewage treatment infrastructure (Kaur et al. 2026). Wyoming, with its extensive ranching operations and decentralized waste management, faces unique water quality challenges. Recent assessments of Fish and Flat Creeks near Jackson, a gateway to the US’s oldest national park and a popular tourist spot, have reported E. coli levels exceeding recreational water standards, raising concerns about waterborne transmission of AMR (Wyoming Department of Environmental Quality 2017).
Traditionally, fecal indicator bacteria such as E. coli and Enterococcus spp. have been used to assess water quality through culture-based methods. However, these approaches provide limited insight into microbial community structure, resistance mechanisms, and genetic exchange among bacteria (Korajkic et al. 2018). Pairing culture-based approaches with advanced molecular techniques, including next-generation sequencing and metagenomics, provides powerful tools for characterizing microbial diversity, ARG prevalence, and resistance-transmission pathways (Sukhum et al. 2019). We hypothesized that recreational waters may serve as reservoirs of mobile MDR gene clusters in E. coli, reflecting diverse anthropogenic and zoonotic inputs.
Despite growing attention to AMR in clinical, agricultural, and wastewater settings, impaired recreational watersheds remain rarely characterized at the genomic level, and few studies have integrated phenotypic susceptibility testing, whole-genome sequencing, and plasmid reconstruction to resolve where resistance determinants reside and how mobile they are. To our knowledge, this is among the first studies to apply this combined approach to an impaired recreational freshwater system. Therefore, the objectives of this study were to isolate E. coli from freshwater samples and characterize their phenotypic resistance profiles, genomic resistance determinants, and clonal relatedness across sampling sites. We further hypothesized that isolates from this impaired recreational watershed would include genetically diverse lineages carrying ExPEC-like virulence traits and MDR plasmids.
Materials and methods
Study area and sampling locations
Presumptive E. coli isolates (n = 263) were collected in 2021 from the Greys-Hoback watershed in Wyoming. Fourteen sampling sites were chosen along Fish and Flat Creeks, both designated as impaired due to elevated E. coli levels, as well as their tributaries, Cache and Game Creeks. Sites were selected to represent upstream, midstream, and downstream locations across the watershed while maintaining accessibility for consistent sampling.
Sample collection and isolate selection
Surface water samples were collected ∼15 cm below the water surface using sterile 1 l polypropylene bottles and a sampling pole to minimize substrate disturbance. Samples were transported on ice and processed within 24 h, filtered through 0.45 µm membrane filters (Pall Corporation, USA) to capture bacteria. The filtered material was then cultured on commercially available modified membrane-thermotolerant E. coli agar (modified mTEC) (Sigma-Aldrich, USA) for selection of E. coli. For quality control, E. coli (ATCC 11775) was used as a positive control, and Enterococcus faecalis (ATCC 19433) as a negative control. Isolates were confirmed via MALDI-TOF MS (Bruker Microflex LRF, Biotyper version 3.0, USA), using a bacterial test standard (Bruker, USA) as a control to identify the colonies. Fresh, isolated colonies were placed on a target plate, dried, and coated with α-cyano-4-hydroxycinnamic acid matrix (Bruker, USA). Scores ≥ 2.300 indicate species-level identification (Popović et al. 2022).
Antimicrobial susceptibility testing
Antimicrobial susceptibility testing (AST) was assessed by broth microdilution according to Clinical and Laboratory Standards Institute (CLSI) guidelines (M100-ED35:2025) using the Sensititre system (Thermo Fisher Scientific, USA). In brief, well-isolated overnight colonies were suspended in sterile 0.85% saline and adjusted to a 0.5 McFarland standard using a nephelometer (CrystalSpec). This suspension was then diluted in Mueller-Hinton broth (Thermo Fisher Scientific) to achieve a final inoculum density of approximately 5 × 105 CFU ml−1 in accordance with CLSI guidelines.
Pre-coated antibiotic plates (CMV3GANF, Thermo Fisher Scientific, USA) were inoculated with 50 μl of suspension and incubated at 37°C for 18–20 h. Minimum inhibitory concentrations (MICs) were determined using SWIN software (v3.3) (Thermo Fisher Scientific, 2024, USA) and interpreted in accordance with CLSI standards (EM100 Connect—CLSI M100 ED35:2025,n.d.). The antimicrobials tested and their respective concentration ranges (µg ml−1) were gentamicin (0.25–16), streptomycin (2–64), amoxicillin-clavulanic acid (1/0.5–32/16), ceftriaxone (0.25–64), ceftiofur (0.12–8), cefoxitin (0.5–32), sulfisoxazole (16–256), trimethoprim-sulfamethoxazole (0.12/2.38–4/76), azithromycin (0.12–16), ampicillin (1–32), chloramphenicol (2–32), ciprofloxacin (0.015–4), nalidixic acid (0.5–32), and tetracycline (4–32).
DNA Extraction and Library Preparation
Escherichia coli isolates exhibiting phenotypic resistance to at least one of the tested antimicrobials were selected for WGS, as described previously (Ghorbani Tajani et al. 2024). In brief, cultures were grown on brain heart infusion agar (Oxoid, USA) at 37°C for ∼18 h. DNA extraction was performed using the DNeasy Blood & Tissue Kit on a QIAcube system (Qiagen) at the Wyoming Public Health Laboratory (Cheyenne, Wyoming), following the manufacturer’s instructions. Libraries were prepared using the NEBNext Ultra DNA Library Prep Kit (NEB, USA). DNA was sonicated to an average fragment size of 350 bp, end-repaired, A-tailed, adapter-ligated, and PCR-amplified with five cycles of indexing. Library quality was assessed using an Agilent 2100 Bioanalyzer, and DNA concentration was measured with a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, USA).
Whole Genome Sequencing and Bioinformatics
Sequencing was performed on an Illumina NovaSeq 6000 platform using 2 × 150-cycle paired-end reads. Raw sequencing reads were evaluated as described previously (Ghorbani Tajani et al. 2025). Briefly, reads were assessed using FastQC (v0.12.0), and low-quality bases and adapters were trimmed using Trimmomatic (v0.36). High-quality reads were assembled de novo using SPAdes (v3.15.5), and genome annotation was performed using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP; v6.8).
ARGs were identified using ABRicate (v1.0.1), and plasmid replicons using PlasmidFinder (v2.0.1). Multilocus Sequence Typing (MLST) via MLST software (v2.0.9), and core genome MLST (cgMLST) analysis conducted with cgMLSTFinder (v1.2). Virulence factors and pathogenic potential were assessed using VirulenceFinder (v2.0) and PathogenFinder (v1.1), respectively. Additional resistance gene profiling was performed using ResFinder (v4.6.0) with the Center for Genomic Epidemiology (CGE) Database (https://www.genomicepidemiology.org/services/). The phylogenetic relationships along with fimH typing were inferred using EnteroBase (v.1.2.0) (https://enterobase.warwick.ac.uk/). Phylogroup assignments were also obtained through EnteroBase and incorporated into the phylogenetic tree to characterize the broader phylogenetic distribution of the isolates.
Draft genome assemblies were screened using MobileElementFinder (v1.0.3) (using the CGE server) to identify mobile genetic elements at the contig level. Based on this analysis, three isolates (S5, S7, and S30) were selected because they showed co-occurrence of plasmid replicons and AMR genes on the same contigs.
Selected assemblies were further analyzed using MOB-suite (mob_recon and mob_typer; v3.1.9) to assign contigs to putative plasmid clusters and predict plasmid mobility based on replicon content, relaxase type, mating pair formation (MPF) system, and oriT-associated features. AMR genes were further identified by screening assemblies and plasmid-associated contigs against the NCBI AMR database using ABRicate (v1.0.1). Resistance genes were considered plasmid-associated only when located on contigs classified as plasmid-derived by MOB-suite. Because plasmid reconstruction from short-read assemblies can be affected by fragmentation and repetitive sequences, results were interpreted conservatively, and ambiguous contigs were not classified as definitively plasmid-borne.
Data Analysis
Statistical and bioinformatic analyses were conducted to characterize AMR, plasmid content, and virulence gene profiles among E. coli isolates. Genomic data were structured as binary presence/absence matrices for ARGs, plasmid replicons, and virulence-associated genes using Python (v3.12) and the pandas library. Prevalence was calculated as the proportion of isolates carrying each genetic feature.
Comparisons of gene prevalence were performed using Fisher’s exact test, appropriate for small sample sizes and categorical data. To account for multiple comparisons, the Bonferroni correction was applied, and adjusted P-values < 0.05 were considered statistically significant.
All statistical analyses were conducted on the subset of phenotypically resistant isolates selected for genomic characterization and therefore do not represent the full E. coli population recovered from the watershed.
Results
Multidrug-resistant E. coli prevalence in a recreational freshwater system
A total of 218 isolates were confirmed as E. coli by MALDI-TOF MS (Bruker, USA), providing the pool from which antibiotic-resistant isolates were selected. AST identified 20 (∼9%) isolates with clinically relevant resistance profiles (Isolates exhibiting resistance to at least one antimicrobial were considered AMR). The resistant isolates were recovered from multiple sampling locations across the watershed, indicating a broad spatial distribution rather than restriction to a single sampling location. Additionally, six isolates (S5, S6, S24, S26, S29, and S30) were found to be MDR, defined as resistance to three or more antimicrobial classes (Fig. 1). Sampling dates for the 20 phenotypically characterized isolates spanned the 2021 recreation season, from early spring (March) through late September.
Figure 1.
(a) Antibiotic susceptibility profiles of E. coli isolates (n = 20) collected from freshwater in the Greys-Hoback watershed. The heatmap shows the susceptibility of individual isolates (y-axis) to a panel of 14 antibiotics (x-axis), categorized as susceptible, intermediate, or resistant according to established clinical breakpoints. Visualization was generated in R (v4.5.2). (b) Prevalence (%) of intermediate and resistant isolates across tested antimicrobials. Resistance (R) and intermediate (I) were determined based on AST results for E. coli isolates (n = 20). Visualization was generated in R (v4.5.2).
Antibiotic resistance patterns in E. coli isolates
Among these 20 resistant isolates, tetracycline resistance was most frequent (16/20, 80%), followed by streptomycin (10/20, 50%). Interestingly, isolate S21 was the only E. coli found to be resistant to cefoxitin, while isolate S3 displayed intermediate resistance to this second-generation cephalosporin. All isolates remained susceptible to third-generation cephalosporins and several other clinically important agents (Fig. 1), indicating that resistance was largely confined to older, commonly used antibiotic classes.
Genomic characterization of MDR and virulence in freshwater E. coli
Genomic and statistical analyses were restricted to the 20 phenotypically resistant E. coli isolates selected for WGS. The blaEC gene family was detected in all 20 genomes (100%). The blaEC-18 allele was the most common (13/20, 65%) and was significantly more prevalent than other detected blaEC alleles after Bonferroni correction (adjusted P < 0.05, Table 1, Fig. 2a).
Table 1.
Genomic characteristics of antibiotic-resistant E. coli isolates (n = 20) recovered from freshwater samples of the Greys-Hoback watershed.
| Isolates | AMR Gene(s) | Plasmid(s) | Virulence Genes | Pathogenicity (%) | GPS Coordinates | Accession Number |
|---|---|---|---|---|---|---|
| S1 | bla EC-19 | IncFIB(AP001918), IncFII | AslA, aalG, anr, astA, chuA, csgA, faeC, faeD, faeE, faeF, faeH, fdeC, fimH, fyuA, gad, irp2, kpsE, kpsMII, nlpI, ompT, papA_F1651A, papC, shiA, terC, traT, usp, yehA, yehB, yehC, yehD, yfcV, yghJ | Yes (93.7%) | 43.393935°, −110.727465° | JBIEAL000000000 |
| S2 | bla EC-18, tet(C) | IncFIB(AP001918), IncFII(pSE11), IncI1-I(Alpha), IncX1 | AslA, F17A, F17C, F17D, F17G, cdt-IIIB, chuA, cnf2, csgA, eilA, eltIIAB-c1, espY2, faeC, faeD, faeE, faeF, faeH, faeI, faeJ, fdeC, fimH, gad, hlyE, iss, iucC, iutA, nlpI, ompT, terC, traT, yehB | Yes (92.9%) | 43.460454°, −110.864262° | JBIEAK000000000 |
| S3 | bla EC-18 | IncFIB(pB171), IncFII(pCoo) | csgA, fdeC, fimH, gad, hlyE, lpfA, nlpI, ompT, terC, traJ, traT, yehA, yehB, yehC, yehD, yghJ | Yes (94.6%) | 43.523742°, −110.863134° | JBIEAJ000000000 |
| S4 | bla EC-18, tet(B) | IncFIA, IncFIB(AP001918), IncFIB(pB171), IncFII(pCoo) | astA, csgA, fdeC, fimH, gad, hlyE, hra, iss, lpfA, nlpI, ompT, shiB, terC, traT, yehA, yehB, yehC, yehD | Yes (93.5%) | 43.460454°, −110.864262° | JBIEAI000000000 |
| S5 | bla EC-18 , floR, tet(A), aph(6)-Id, aph(3′’)-Ib, sul2, bla TEM-150 | IncFIA, IncFIB(AP001918), IncFIC(FII) | anr, csgA, fdeC, fimH, gad, hlyE, iss, lpfA, nlpI, ompT, terC, traJ, traT, yehA, yehB, yehC, yehD | Yes (94.1%) | 43.49565°, −110.871933° | JBIEAH000000000 |
| S6 | bla EC-18 | No hits found | F17A, F17C, F17D, F17G, astA, csgA, fdeC, fimH, gad, hlyE, hra, iss, lpfA, nlpI, ompT, shiB, terC, yehA, yehB, yehC, yehD, yghJ | Yes (93.7%) | 43.39056°, −110.748393° | JBIEAG000000000 |
| S7 | tet(A), bla EC-18 | IncFIB(AP001918), IncFII(pCoo), IncI1-I(Alpha) | csgA, fdeC, fimH, gad, hlyE, lpfA, nlpI, ompT, terC, yehA, yehB, yehC, yehD, yghJ | Yes (94.3%) | 43.49565°, −110.871933° | JBIDZM000000000 |
| S9 | bla EC-18, tet(B) | IncFIB(AP001918), IncY | csgA, fdeC, fimH, gad, hlyE, lpfA, nlpI, ompT, terC, traT, yehA, yehB, yehC, yehD, yghJ | Yes (94.0%) | 43.48778°, −110.763211° | JBIEAD000000000 |
| S10 | aph(3′’)-Ib, aph(6)-Id, sul2, bla EC-18 , tet(B) | IncFIB(AP001918), IncFII(pCoo) | aamR, csgA, fdeC, fimH, gad, hlyE, hra, iss, lpfA, nlpI, ompT, papC, shiA, shiB, terC, tia, traT, yehA, yehB, yehC, yehD, yghJ | Yes (93.8%) | 43.523742°, −110.863134° | JBIEAC000000000 |
| S19 | bla EC-18 | IncFII, IncX1 | AslA, F17A, F17C, F17D, air, astA, chuA, csgA, eilA, fdeC, fimH, gad, hha, hlyE, hra, iss, nlpI, shiB, terC, tia, traT, yehA | Yes (92.8%) | 43.460454°, −110.864262° | JBIDZU000000000 |
| S20 | bla EC-18 | IncI2(Delta) | colE2-like, csgA, fdeC, fimH, gad, hlyE, iss, lpfA, nlpI, ompT, terC, yehA, yehB, yehC, yehD, yghJ | Yes (94.4%) | 43.39056°, −110.748393° | JBIDZT000000000 |
| S21 | bla EC-18 , sul2 | IncX1 | F17A, F17C, F17D, csgA, fdeC, fimH, gad, hha, hlyE, hra, nlpI, shiB, terC, yehA, yehB, yehC, yehD, yghJ | Yes (93.5%) | 43.39056°, −110.748393° | JBIDZS000000000 |
| S22 | bla EC-18 , sul2, aph(6)-Id, aph(3′’)-Ib, tet(B) | IncFIB(pB171), IncFII(pCoo) | F17A, F17D, F17G, astA, csgA, fdeC, fimH, fyuA, gad, hha, hlyE, hra, irp2, lpfA, ompT, terC, tia, traJ, traT, yehA, yehB, yehC, yehD, yghJ | Yes (93.6%) | 43.45835°, −110.797119° | JBIDZR000000000 |
| S23 | bla EC-5 , tet(A) | IncFIA, IncFIB(AP001918), IncFIC(FII) | AslA, anr, chuA, csgA, cvaC, hak, etsC, fdeC, fimH, fyuA, gad, hlyF, ibeA, iroN, irp2, iss, iucC, iutA, kpsE, kpsMII_K1, mchF, neuC, nlpI, ompT, sitA, terC, traJ, traT, tsh, usp, vat, yehA, yehB, yehC, yehD, yfcV | Yes (93.9%) | 43.39056°, −110.748393° | JBIDZQ000000000 |
| S24 | bla EC-18 , sul2, aph(3′’)-Ib, aph(6)-Id, tet(A), floR | IncFIB(AP001918), IncFII(pCoo) | F17A, F17C, F17D, F17G, csgA, eltIIAB-c3, fdeC, fimH, fyuA, gad, hha, hlyE, hra, irp2, iss, lpfA, nlpI, shiB, terC, traJ, yehA, yehB, yehC, yehD, yghJ | Yes (93.9%) | 43.523742°, −110.863134° | JBIDZP000000000 |
| S25 | bla EC-8 | Col156, IncFIB(AP001918), IncI1-I(Alpha), IncX1 | AslA, air, chuA, colE2-like, csgA, eilA, faeE, faeF, faeH, faeI, faeJ, fdeC, fimH, gad, hha, hlyE, iss, kpsE, kpsMII, lpfA, nlpI, ompT, papC, terC, yehA, yehB, yehC, yehD, yghJ | Yes (93.6%) | 43.574643°, −110.836912° | JBIDZO000000000 |
| S26 | bla EC-8 | No hits found | csgA, fdeC, fimH, gad, hha, hlyE, hra, lpfA, nlpI, terC, yehA, yehB, yehC, yehD, yghJ | Yes (94.3%) | 43.432652°, −110.792589° | JBIDZN000000000 |
| S28 | bla EC-8 | No hits found | F17A, F17C, F17D, F17G, astA, csgA, fdeC, fimH, fyuA, gad, hha, hlyE, hra, iss, lpfA, nlpI, ompT, terC, tia, yehA, yehB, yehC, yehD, yghJ | Yes (93.8%) | 43.432652°, −110.792589° | JBIDZL000000000 |
| S29 | bla EC-8 | IncFIB(AP001918), IncFII(pSE11) | F17A, F17C, F17D, F17G, aalG, anr, cdt-IIIB, cia, cnf2, csgA, eltIIAB-c2, etsC, faeJ, fdeC, fimH, gad, hlyE, hlyF, iss, iucC, iutA, lpfA, nlpI, ompT, terC, traJ, traT, yehA, yehB, yehC, yehD, yghJ | Yes (94.1%) | 43.39056°, −110.748393° | JBIDZK000000000 |
| S30 | bla EC-8 , floR, tet(A), aph(6)-Id, aph(3′')-Ib, sul2 | IncFIB(AP001918), IncFII(pCoo) | F17A, F17C, F17D, F17G, aalG, anr, cdt-IIIB, cia, cnf2, csgA, eltIIAB-c2, etsC, faeJ, fdeC, fimH, gad, hlyE, hlyF, iss, iucC, iutA, lpfA, nlpI, ompT, terC, traJ, traT, yehA, yehB, yehC, yehD, yghJ | Yes (93.9%) | 43.523742°, −110.863134° | JBIDZJ000000000 |
Figure 2.
Prevalence of (a) Antimicrobial Resistance Genes, (b) Plasmid Replicons, and (c) Virulence Genes among E. coli isolates (n = 20). Prevalence was calculated as the percentage of isolates in which each marker was detected based on whole-genome sequencing data. For each isolate, marker presence was treated as a binary variable (present/absent), and duplicate entries within isolates were removed. Plasmid entries reported as “No hits found” (n = 3) were excluded from analysis. Visualization was generated using Matplotlib in Python (v3.12).
Less common β-lactam determinants included blaTEM-150, which was detected only in isolate S5, as well as additional blaEC variants such as blaEC-19, blaEC-5, and blaEC-8. Non-β-lactam ARGs were more frequently detected, including tetracycline resistance genes (tet(A), tet(B), tet(C)) (10/20, 50%), the sulfonamide resistance gene sul2 (6/20, 30%), aminoglycoside resistance genes aph(3′')-Ib and aph(6)-Id (each 5/20, 25%), and the phenicol resistance gene floR (3/20, 15%) (Table 1, Fig. 2a).
Plasmid replicons were detected in 17 of the 20 (85%) E. coli isolates, with many isolates harboring multiple incompatibility groups. Isolates S6, S26, and S28 lacked any detectable plasmid replicons, whereas isolates S2 and S25 both carried the greatest diversity of plasmid types, each harboring four replicon families (IncFIB(AP001918), IncFII[pSE11 or pCoo], IncI1-I(Alpha), and IncX1 or Col156).
IncF-type replicons predominated with IncFIB(AP001918) detected in 12/20 isolates (60%) and IncFII variants (including IncFII(pCoo) and IncFII(pSE11)) also present in 11/20 isolates (55%). Additional plasmid backbones included IncX1 and IncFIA (4/20, 20%; 3/20, 15% respectively), with IncI1-I(Alpha) (3/20, 15%), IncI2(Delta) (1/20, 5%), IncY (1/20, 5%), and the small Col156 replicon (1/20, 5%). These findings indicate that multi-replicon IncF plasmids constitute the dominant vehicles for ARG carriage in this E. coli population (Table 1, Fig. 2b). Notably, S5 combined the highest ARG count with three IncF plasmids (Table 1, Fig. 2b), highlighting this isolate as a notable MDR, plasmid-rich genotype.
All E. coli isolates harbored multiple virulence-associated genes (20/20, 100%) (Table 1, Fig. 2c). Most isolates (including S3-S7, S9, S10, S19-S22, S24-S26, S28-S30) carried core adherence and serum-resistance factors such as csgA, fdeC, fimH, gad, hlyE, lpfA, nlpI, ompT, terC, and the tellurite resistance locus yehA-yehD. Several isolates (S1, S19, S21, S22, S23, S24, S25, S28–S30) showed expanded ExPEC- and ETEC-associated repertoires, encoding combinations of fyuA and/or irp2 (yersiniabactin siderophore system), kpsE/kpsMII (capsule), papA/papC (P fimbriae), eltIIAB variants (LT toxin), tsh, vat, ibeA, iroN, and multiple F17 fimbrial variants (F17A, F17C, F17D, F17G). Toxin and intestinal pathogenicity markers such as astA, air, cvaC, colE2-like, cdt-IIIB, elt-IIAB, eilA, and hlyF were detected in subsets of isolates (notably S2, S19, S22-S25, S29, and S30), indicating that several strains combine extraintestinal and intestinal virulence traits. Eleven virulence genes (csgA, fdeC, fimH, gad, terC, nlpI, yehA, yehB, hlyE, yehC, and yehD) were significantly more prevalent than low-frequency virulence markers after Bonferroni correction (adjusted P < 0.05).
Many additional virulence factors occurred at low frequencies and collectively contributed to the overall heterogeneity in virulence-gene content, and all 20 isolates were predicted to be potentially pathogenic (Table 1).
Genetic diversity of clinically relevant E. coli lineages
Core genome multilocus sequence typing (cgMLST; EnteroBase cgMLST v1) combined with HierCC clustering revealed substantial genetic diversity among the 20 E. coli isolates. A GrapeTree phylogeny constructed using the NINJA neighbor-joining algorithm revealed multiple sequence types (STs) and distinct clonal groupings (Fig. 3). HierCC analysis further indicated both closely related clusters and distantly related isolates, supporting the coexistence of diverse lineages within this system. Phylogroup classification further supported the observed genetic diversity, with isolates distributed across multiple phylogroups, including A, B1, B2, D, and E, as well as an ABD assignment (Fig. 3).
Figure 3.
Phylogenetic tree illustrating the relationships among freshwater E. coli isolates (n = 20) from the Greys-Hoback watershed, inferred using EnteroBase version 1.2.0 cgMLST (cgMLST V1 + HierCC V1); based on NINJA NJ algorithm and visualized and annotated in Interactive Tree of Life (iTOL) version 7.2.2. The outer annotation rings indicate the sequence type (ST) and phylogroup assigned to each isolate.
Several STs belonged to recognized clonal complexes including the ST155 complex (S4, S6, S22, S28), ST156 complex (S24, S30), ST446 complex (S5, S29), and ST86 complex (S7). Other isolates were classified as singletons without an assigned complex, underscoring the presence of both common and rare lineages in this recreational system.
Notably, isolate S1 (ST646; serotype O4: H7; fimH5) carried an expanded ExPEC-associated virulence repertoire, including the yersiniabactin siderophore system (fyuA, irp2), group 2 capsule loci (kpsE, kpsMII), P fimbriae (papA, papC), the uropathogenic-specific protein gene usp, and the heme-acquisition gene chuA. Together, these findings suggest that this system harbors lineages with both enhanced virulence potential and a substantial mobile resistome.
Additionally, serotype prediction based on O- and H-antigen loci was performed using EnteroBase (Table 2), identifying a broad range of O: H combinations, including O4: H7, O7: H21, O8: H7, O17: H41, O23: H25, O25: H21, O38: H39, O84: H14, O85: H8, O87: H7, O98: H41, O134: H31, O153: H25, O27: H21, O156: H21, and O184: H48, as well as two isolates (S4 and S7) with only H-antigen assignments (H21 and H7, respectively).
Table 2.
Genetic profiles of E. coli isolates (n = 20) from the Greys-Hoback watershed, including MLST sequence types, cgMLST allele numbers, O: H serotypes, and fimH alleles, highlighting genetic diversity.
| Isolates | MLST | cgMLST | Serotype | FimTyper |
|---|---|---|---|---|
| S1 | 646 | 166 067 | O4:H7 | fimH5 |
| S2 | 1087 | 89 388 | O98:H41 | - |
| S3 | 442 | 105 373 | O156:H21 | fimH32 |
| S4 | 155 | 114 842 | H21 | fimH121 |
| S5 | 1704 | 125 114 | O87:H7 | fimH31 |
| S6 | 56 | 76 819 | O25:H21 | fimH54 |
| S7 | 1308 | 4609 | H7 | fimH32 |
| S9 | 196 | 196 780 | O8:H7 | fimH87 |
| S10 | 937 | 55 523 | O184:H48 | fimH38 |
| S19 | 6975 | 90 215 | O38:H39 | fimH30 |
| S20 | 2144 | 43 895 | O84:H14 | - |
| S21 | 336 | 98 481 | O27:H21 | fimH297 |
| S22 | 58 | 101 148 | O7:H21 | fimH32 |
| S23 | 569 | 14 556 | O134:H31 | fimH5 |
| S24 | 156 | 138 645 | O23:H25 | fimH38 |
| S25 | 720 | 89 793 | O17:H41 | fimH65 |
| S26 | 2522 | 132 463 | O85:H8 | fimH38 |
| S28 | 56 | 104 712 | O153:H25 | fimH54 |
| S29 | 666 | 4530 | O8:H7 | fimH31 |
| S30 | 156 | 138 645 | O23:H25 | fimH38 |
Plasmid-associated AMR in selected isolates
Contig-level analysis of isolates S5, S7, and S30 refined the genomic localization of AMR determinants and revealed distinct plasmid-associated patterns. In S5, multiple resistance genes, including floR, tet(A), aph(6)-Id, aph(3′')-Ib, sul2, and blaTEM-150, were localized to a plasmid cluster assigned by MOB-suite to an IncF-type plasmid with conjugative features. This finding supports the presence of an MDR plasmid in S5.
In S30, five resistance genes, floR, tet(A), aph(6)-Id, aph(3′')-Ib, and sul2, co-localized on a single plasmid-associated contig. A separate contig carrying blaEC-18 was classified as chromosomal, indicating that the β-lactamase gene was not part of the plasmid-associated MDR region. Although the S30 plasmid lacked identifiable mobility-associated features and was predicted to be non-mobilizable, it still represented a discrete plasmid-associated resistance element.
In contrast, S7 yielded conflicting results between analytical approaches. While plasmid replicon screening detected the presence of plasmid-associated sequences, MOB-suite classified the tet(A)-containing contig as chromosomal and identified a separate plasmid contig lacking AMR genes. Therefore, S7 was considered ambiguous and was not retained as a confirmed plasmid-associated AMR case.
Discussion
The findings in this study support the concept that recreational freshwater environments can serve as reservoirs of potentially mobile MDR gene clusters. Only 20 of 218 confirmed isolates displayed clinically relevant resistance phenotypes, and six were classified as MDR. This classification followed established criteria for interpreting MDR (resistant to three or more antimicrobial classes) (Magiorakos et al. 2012). Nevertheless, these findings remain important because the watershed is actively used for recreation and lies within a broader One Health interface where human activity, wildlife, livestock, and surface water intersect. In this setting, E. coli serves primarily as an indicator of fecal contamination rather than a direct measure of pathogen risk. However, previous work has shown that indicator-positive waters can contain isolates with clinically relevant resistance and virulence traits (Franz et al. 2015, Korajkic et al. 2018). These results provide genomic evidence that even rural freshwater systems can serve as reservoirs for resistant E. coli, with potential relevance to human exposure.
Resistance in this watershed was dominated by tetracycline and streptomycin, with lower frequencies of chloramphenicol and ampicillin resistance and no phenotypic resistance to ceftriaxone or ceftiofur (Fig. 1). This pattern suggests that resistance within the watershed is primarily driven by widely distributed legacy, non-ESBL determinants of resistance rather than by high-risk ESBL phenotypes commonly associated with clinical settings. Similar patterns have been reported in aquatic environments, where tetracycline resistance is often the predominant phenotype, especially in systems influenced by mixed anthropogenic and animal inputs (Haberecht et al. 2019, Kaur et al. 2026). At the watershed scale, this pattern is consistent with the integration of inputs from agricultural runoff, wildlife, and human-associated sources, allowing resistance determinants from different origins to co-occur (Kraemer et al. 2022).
WGS provided important context for interpreting these patterns. Although blaEC variants were detected across the sequenced isolates, these findings should be interpreted alongside the observed susceptibility to third-generation cephalosporins and the absence of common ESBL genes (Fig. 2a). Notably, phenotypic resistance exceeded the detected genotypic determinants for several agents: 16 of 20 isolates were tetracycline-resistant but only 10 carried a tet gene, and 10 were streptomycin-resistant despite only 5 carrying aph(3′')-Ib/aph(6)-Id, suggesting contributions from efflux, chromosomal mutations, and more (Darby et al. 2023). Together, these findings highlight the importance of integrating phenotypic and genomic data for accurate interpretation.
Overall, these genomes do not indicate a strong dominance of clinically critical cephalosporin resistance. Instead, the consistent signal was the repeated detection of non-β-lactam resistance genes, including tet, sul2, aph(3′')-Ib, aph(6)-Id, and floR. The co-occurrence, particularly in MDR isolates, indicates a mobile resistome consistent with environmental and animal-associated E. coli populations (Fig. 2a). These findings support the idea that recreational waters may function not only as reservoirs of resistant bacteria but also as environments where resistance genes persist and may exchange.
To better understand the genomic context of resistance, plasmid-associated elements were analyzed in detail. Most isolates carried plasmid replicons, and IncFIB and IncFII were the dominant backbones (Fig. 2b). IncF plasmids are widely recognized as important vehicles for the persistence and dissemination of AMR and virulence traits in E. coli and related Enterobacterales (Pitout and Chen 2023). Their presence does not demonstrate active horizontal transfer within the watershed; however, they indicate that many isolates harbor genomic architectures that are compatible with the maintenance and spread of adaptive traits. The presence of multi-replicon profiles and the clustering of multiple resistance genes in isolates such as S5, S24, and S30 further support the presence of a mobile gene pool within this system.
The contig-level analysis suggests that plasmids contribute substantially to resistance gene carriage in this collection, though its contribution varied across isolates. In isolate S5, the co-localization of floR, tet(A), aph(6)-Id, aph(3′')-Ib, sul2, and blaTEM-150 genes on a conjugative IncF-type plasmid is of certain concern, since physical linkage of these determinants on a mobile element selects for co-selection and horizontal transfer; exposure to any one agent could sustain the entire cassette. IncF plasmids are well-recognized vehicles for MDR in Enterobacteriaceae (Stein et al. 2024), and our findings are consistent with this role. The recurrence of the same floR, tet(A), aph(6)-Id, aph(3′')-Ib, sul2 genes in S30 potentially indicates a conserved resistance region circulating among these isolates. However, it remained limited to a plasmid contig while the intrinsic blaEC-18 stayed chromosomal. This partial segregation usefully distinguishes mobile, potentially transferable resistance from determinants that may be tied to the chromosome and therefore less likely to disseminate. On the contrary, isolate S7 was harder to decipher as plasmid replicons were detected, yet its AMR genes assembled onto chromosomal contigs, leaving the association unresolved. We consider this most likely an artifact of assembly fragmentation rather than a true absence of plasmid-borne resistance, a recognized limitation of reconstructing repetitive plasmid sequences from short reads, as described previously (Arredondo-Alonso et al. 2017, Teixeira et al. 2025).
These findings highlight both the contribution of plasmids to resistance gene clustering and the limitations of plasmid reconstruction using short-read sequencing approaches (Arredondo-Alonso et al. 2017). Importantly, the detection of plasmid-associated MDR regions in isolates from recreational waters underscores the potential for environmental dissemination of clinically relevant resistance determinants beyond traditional clinical or agricultural settings.
In addition to AMR, all sequenced isolates carried multiple virulence-associated genes, and several encoded combinations consistent with both extraintestinal and enterotoxigenic potential. This aligns with previous studies showing that environmental E. coli can harbor overlapping resistance and virulence traits, including in surface water and wastewater-impacted systems (Franz et al. 2015, Gomes et al. 2024). However, these findings still need to be interpreted with caution, as genome-based pathogenicity predictions and virulence gene profiles reflect potential rather than confirmed clinical behavior, as they do not account for gene expression, infective dose, or host susceptibility. Accordingly, these predictions should be considered indicative rather than definitive measures of pathogenic risk.
The serotypic heterogeneity observed here is consistent with repeated introduction of resistant E. coli from multiple fecal inputs rather than the expansion of a single dominant clone. Broad serotype and lineage diversity is a recurrent feature of recreational and mixed-use surface waters, where E. coli populations commonly include numerous co-occurring serotypes and phylogenetic backgrounds and, in some settings, substantial MDR burdens (Nadya et al. 2016, Haberecht et al. 2019). Because O: H antigen loci can occur on different chromosomal backgrounds, serotype data alone do not establish source attribution; however, when interpreted together with MLST/cgMLST, such heterogeneity provides complementary evidence of population-level diversity and repeated inputs rather than recent clonal expansion (Ingle et al. 2016). This interpretation is further supported by aquatic AMR studies showing that resistant E. coli are phylogenetically dispersed and carry diverse resistome across water systems, again arguing against a single dominant resistant lineage (Haberecht et al. 2019, Ogura et al. 2020).
Phylogenetic analyses further demonstrated that the resistant E. coli population in this watershed is highly diverse (Fig. 3). MLST, cgMLST, serotyping, and fimH typing revealed multiple lineages, including both clustered and genetically distinct isolates, consistent with repeated environmental introduction rather than clonal expansion. The distribution of AMR isolates across multiple E. coli phylogroups further supports the presence of a genetically heterogeneous population within the watershed and is consistent with potential contributions from multiple ecological sources. Several isolates belonged to recognized clonal complexes, including ST155, ST156, ST446, and ST86, while others were singletons, reflecting a mix of common and rare lineages. This pattern is consistent with the ecology of mixed-use watersheds, where bacterial inputs come from wildlife, livestock, septic systems, and human activity. Similar heterogeneity has been reported in other environmental water studies, including systems influenced by wastewater and landscape-scale runoff (Haberecht et al. 2019, Kaur et al. 2026). Accordingly, lineage diversity may be as informative as overall resistance prevalence, indicating a distributed and continuously replenished AMR burden.
Several limitations should be considered when interpreting these findings. First, genomic analyses were limited to 20 phenotypically resistant isolates and therefore do not represent the full E. coli population present in the watershed. Second, the study design does not allow direct source attribution; incorporating microbial source tracking or land-use data would strengthen ecological interpretation. Third, plasmid reconstruction from short-read assemblies remains challenging, and detection of plasmid replicons does not confirm plasmid completeness or transferability. Finally, pathogenicity predictions should be interpreted as screening-level evidence rather than definitive indicators of human health risk.
Despite these limitations, this study provides a framework for linking routine water-quality monitoring with genomic surveillance of AMR. Incorporating AST, targeted WGS, and plasmid analysis into recreational water monitoring could improve early detection of high-risk resistance profiles and support overall risk assessment. More broadly, these findings support the inclusion of recreational freshwater systems in One Health AMR surveillance efforts, particularly in regions where environmental, agricultural, and human factors intersect.
Collectively, these findings highlight recreational waters as under-recognized reservoirs of potentially mobile MDR gene clusters with potential zoonotic relevance, emphasizing their importance within One Health AMR surveillance frameworks.
Acknowledgements
The authors would like to acknowledge Wyoming Public Health Laboratory, Cheyenne, Wyoming, for their work on whole-genome sequencing. The authors also wish to acknowledge the valuable contributions of Dr. Sarah Collins, whose efforts in conceptualization, project development, and field coordination were instrumental to this work. Her contributions to study conceptualization, project development, and field coordination were completed prior to final manuscript preparation. Ethical approval was not required for this study because it involved the collection and analysis of environmental water samples and did not involve human participants, animals, or identifiable personal data.
Contributor Information
Anahita Ghorbani Tajani, Department of Animal Science, University of Wyoming, Laramie, WY 82071, United States.
Aniket Sharma, Department of Animal Science, University of Wyoming, Laramie, WY 82071, United States.
Kelsey C Ruehling, Department of Zoology and Physiology, University of Wyoming, Laramie, WY 82071, United States.
Bledar Bisha, Department of Animal Science, University of Wyoming, Laramie, WY 82071, United States.
Author contributions
Anahita Ghorbani Tajani (Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing), Aniket Sharma (Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing), Kelsey Ruehling (Data curation, Formal Analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing), Bledar Bisha (Conceptualization, Data curation, Formal Analysis, Funding acquisition, Methodology, Project administration, Writing – original draft, Writing – review & editing)
Conflicts of interest
No conflict of interest declared.
Funding
This research received support from the Wyoming Water Research Program grant, the National Institute of Food and Agriculture, United States Department of Agriculture (Grant Number WYO-618-20), and was also funded by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health (Grant Number 2P20GM103432).
Data availability
Whole-genome sequencing data have been deposited in NCBI under BioProject PRJNA1146023. Accession numbers as listed in Table 1.
References
- Antimicrobial Resistance Collaborators . Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet Lond Engl. 2022;399:629–55. 10.1016/S0140-6736(21)02724-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Arredondo-Alonso S, Willems RJ, Schaik Wvan et al. On the (im)possibility of reconstructing plasmids from whole-genome short-read sequencing data. Microb Genomics. 2017;3:e000128. 10.1099/mgen.0.000128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Boidya P, Ghorbani Tajani A, Blouin N et al. Draft genome sequence of multidrug-resistant Raoultella ornithinolytica isolated from a wastewater treatment plant in Laramie, Wyoming. Microbiol Resour Announc. 2025;15:e01051–25. 10.1128/mra.01051-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chandross-Cohen T, Ebersole M, Rahman N et al. Whole-genome sequence and phenotypic characteristics of Bacillus cereus strain IM8 isolated from the upper Fremont Glacier in Wyoming. Microbiol Resour Announc. 2026;0:e01260–25. 10.1128/mra.01260-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Darby EM, Trampari E, Siasat P et al. Molecular mechanisms of antibiotic resistance revisited. Nat Rev Micro. 2023;21:280–95. 10.1038/s41579-022-00820-y [DOI] [PubMed] [Google Scholar]
- Clinical and Laboratory Standards Institute . M100: Performance Standards for Antimicrobial Susceptibility Testing, EM100 Connect. 35th edn. Clinical and Laboratory Standards Institute, 2025. https://em100.edaptivedocs.net/GetDoc.aspx?doc=CLSI%20M100%20ED35:2025&xormat=SPDF&src=BB (24 Nov. 2025, date last accessed). [Google Scholar]
- Franz E, Veenman C, Hoek AHAMvan et al. Pathogenic Escherichia coli producing extended-spectrum β-Lactamases isolated from surface water and wastewater. Sci Rep. 2015;5:14372. 10.1038/srep14372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ghorbani Tajani A, Hamidi A, Sharma A et al. Draft genome sequence of multidrug-resistant Escherichia coli isolated from retail chicken liver in Kosovo. Microbiol Resour Announc. 2026;0:e01033–25. 10.1128/mra.01033-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ghorbani Tajani A, Sharma A, Blouin N et al. Genome sequence, antibiotic resistance genes, and plasmids in a monophasic variant of Salmonella typhimurium isolated from retail pork. Microbiol Resour Announc. 2024;13:e0075423. 10.1128/mra.00754-23 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ghorbani Tajani A, Sharma A, Ruehling K et al. Molecular characterization of Escherichia coli isolate from the Greys-Hoback Watershed, Wyoming. Microbiol Resour Announc. 2025;0:e01187–24. 10.1128/mra.01187-24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gomes L, Bordalo AA, Machado A. Characterization of Escherichia coli isolates in recreational waters: implications for public health and one health approach. Water. 2024;16:2695. 10.3390/w16182695 [DOI] [Google Scholar]
- Haberecht HB, Nealon NJ, Gilliland JR et al. Antimicrobial-resistant Escherichia coli from environmental waters in Northern Colorado. J Environ Public Health. 2019;2019:3862949. 10.1155/2019/3862949 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ingle DJ, Valcanis M, Kuzevski A et al. In silico serotyping of E. coli from short read data identifies limited novel O-loci but extensive diversity of O:h serotype combinations within and between pathogenic lineages. Microb Genomics. 2016;2:e000064. 10.1099/mgen.0.000064 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kaur H, Wang B, Ruehling K et al. Identification and characterization of Escherichia coli isolated from the surface water of the Laramie River, Wyoming. J Water Health. 2026;24:93–108. 10.2166/wh.2026.091 [DOI] [PubMed] [Google Scholar]
- Wyoming Department of Environmental Quality . DEQ releases results of E. coli sampling on Flat Creek, Fish Creek, and Snake River. Wyo Dep Environ Qual. 2017. http://deq.wyoming.gov/2017/08/deq-releases-results-of-e-coli-sampling-on-flat-creek-fish-creek-and-snake-river/ (28 Sep 2026, date last accessed). [Google Scholar]
- Korajkic A, McMinn BR, Harwood VJ. Relationships between microbial indicators and pathogens in recreational water settings. Int J Environ Res Public Health. 2018;15:art. 12. 10.3390/ijerph15122842 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kraemer SA, Barbosa da Costa N, Oliva A et al. A resistome survey across hundreds of freshwater bacterial communities reveals the impacts of veterinary and human antibiotics use. Front Microbiol. 2022;13:995418. 10.3389/fmicb.2022.995418 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Magiorakos AP, Srinivasan A, Carey RB et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect Off Publ Eur Soc Clin Microbiol Infect Dis. 2012;18:268–81. 10.1111/j.1469-0691.2011.03570.x [DOI] [PubMed] [Google Scholar]
- Nadya S, Delaquis P, Chen J et al. Phenotypic and genotypic characteristics of Shiga Toxin-Producing Escherichia coli isolated from surface waters and sediments in a Canadian Urban-Agricultural Landscape. Front Cell Infect Microbiol. 2016;6:36. 10.3389/fcimb.2016.00036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ogura Y, Ueda T, Nukazawa K et al. The level of antimicrobial resistance of sewage isolates is higher than that of river isolates in different Escherichia coli lineages. Sci Rep. 2020;10:17880. 10.1038/s41598-020-75065-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perry JA, Westman EL, Wright GD. The antibiotic resistome: what’s new?. Curr Opin Microbiol Antimicrobials. 2014;21:45–50. 10.1016/j.mib.2014.09.002 [DOI] [PubMed] [Google Scholar]
- Pitout JDD, Chen L. The significance of epidemic plasmids in the success of multidrug-resistant drug pandemic extraintestinal pathogenic Escherichia coli. Infect Dis Ther. 2023;12:1029–41. 10.1007/s40121-023-00791-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Popović NT, Kepec S, Kazazić SP et al. Identification of environmental aquatic bacteria by mass spectrometry supported by biochemical differentiation. PLoS One. 2022;17:e0269423. 10.1371/journal.pone.0269423 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stein M, Brinks E, Loop J et al. Antibiotic resistance plasmids in Enterobacteriaceae isolated from fresh produce in northern Germany. Microbiol Spectr. 2024;12:e00361–24. 10.1128/spectrum.00361-24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sukhum KV, Diorio-Toth L, Dantas G. Genomic and metagenomic approaches for predictive surveillance of emerging pathogens and antibiotic resistance. Clin Pharmacol Ther. 2019;106:512–24. 10.1002/cpt.1535 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Teixeira M, Souque C, Worby CJ et al. Circling in on plasmids: benchmarking plasmid detection and reconstruction tools for short-read data from diverse species. Brief Bioinform. 2025;26:bbaf589. 10.1093/bib/bbaf589 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Whole-genome sequencing data have been deposited in NCBI under BioProject PRJNA1146023. Accession numbers as listed in Table 1.