Supplementary data are available at Bioinformatics online.

Escherichia coli that are present in the rivers are mostly brought by human and animal feces. Contamination occurs mostly through wastewater treatment plant (WWTP) outflows and field amendment with sewage sludge or manure. However, the survival of these isolates in river-associated wetlands remains unknown. Here, we assessed E. coli population structure in low-anthropized wetlands located along three floodplains to identify the major source of contamination of wetlands, whose functioning is different from the rivers. We retrieved 179 E. coli in water samples collected monthly from 19 sites located in eastern France over 1 year. Phylogroups B1 and B2 were dominant in the E. coli population, while phylogroup A was dominant in isolates resistant to thirdgeneration cephalosporins, which harbored the extended-spectrum β-lactamase (ESBL) encoding genes bla CTX-M-15 and bla CTX-M-27 in half of the cases. The high proportion of isolates from human source can be attributed to WWTP outflows and the spread of sewage sludge. We analyzed the distribution of the isolates belonging to the most human-associated phylogroups (B2 and D) on a phylogenetic tree of the whole species and compared it with that of isolates retrieved from patients and from WWTP outflows. The distribution of the three E. coli populations was similar, suggesting the absence of a specific population in the environment. Our results suggest that a high proportion of E. coli isolates that reach and survive in low-anthropized environments such as wetlands are from human source. To the best of our knowledge, this is the first study assessing E. coli contamination and resistance genes in natural freshwater wetlands.

Next-generation sequencing (NGS) is now widely used in microbiology to explore genome evolution and the structure of pathogen outbreaks. Bioinformatics pipelines readily detect single-nucleotide polymorphisms or short indels. However, bacterial genomes also evolve through the action of small transposable elements called insertion sequences (ISs), which are difficult to detect due to their short length and multiple repetitions throughout the genome. We designed panISa software for the ab initio detection of IS insertions in the genomes of prokaryotes. PanISa has been released as open source software (GPL3) available from https://github.com/bvalot/panISa. In this study, we assessed the utility of this software for evolutionary studies, by reanalysing five published datasets for outbreaks of human major pathogens in which ISs had not been specifically investigated. We reanalysed the raw data from each study, by aligning the reads against reference genomes and running panISa on the alignments. Each hit was automatically curated and IS-related events were validated on the basis of nucleotide sequence similarity, by comparison with the ISFinder database. In Acinetobacter baumannii , the panISa pipeline identified ISAba1 or ISAba125 upstream from the ampC gene, which encodes a cephalosporinase in all third-generation cephalosporin-resistant isolates. In the genomes of Vibrio cholerae isolates, we found that early Haitian isolates had the same ISs as Nepalese isolates, confirming the inferred history of the contamination of this island. In Enterococcus faecalis , panISa identified regions of high plasticity, including a pathogenicity island enriched in IS-related events. The overall distribution of ISs deduced with panISa was consistent with SNP-based phylogenic trees, for all species considered. The role of ISs in pathogen evolution has probably been underestimated due to difficulties detecting these transposable elements. We show here that panISa is a useful addition to the bioinformatics toolbox for analyses of the evolution of bacterial genomes. PanISa will facilitate explorations of the functional impact of ISs and improve our understanding of prokaryote evolution.

No abstract