Remarkable genomic diversity among Escherichia isolates recovered from healthy chickens

Abstract

The genus Escherichia has been extensively studied and it is known to encompass a range of commensal and pathogenic bacteria that primarily inhabit the gastrointestinal tracts of warm-blooded vertebrates. However, the presence of E. coli as a model organism and potential pathogen has diverted attention away from commensal strains and other species in the genus. To investigate the diversity of Escherichia in healthy chickens, we collected fecal samples from antibiotic-free Lohmann Brown layer hens and determined the genome sequences of 100 isolates, 81 of which were indistinguishable at the HC0 level of the Hierarchical Clustering of Core Genome Multi-Locus Sequence Typing scheme. Despite initial selection on CHROMagar Orientation medium, which is considered selective for E. coli, in silico phylotyping and core genome single nucleotide polymorphism analysis revealed the presence of at least one representative of all major clades of Escherichia, except for E. albertii, Shigella, and E. coli phylogroup B2 and cryptic clade I. The most frequent phylogenomic groups were E. coli phylogroups A and B1 and E. ruysiae (clades III and IV). We compiled a collection of reference strains isolated from avian sources (predominantly chicken), representing every Escherichia phylogroup and species, and used it to confirm the phylogeny and diversity of our isolates. Overall, the isolates carried low numbers of the virulence and antibiotic resistance genes typically seen in avian pathogenic E. coli. Notably, the clades not recovered are ones that have been most strongly associated with virulence by other studies.

Keywords: Chickens; Commensal; Cryptic clades; Escherichia; Genomic diversity; Phylogenomics.

Figure 1. Summary of the major stages of sample collection, processing and data analysis and the primary programs and databases used.

Figure 2. Core SNP maximum likelihood tree of the final 81 Escherichia isolates, including reference sequences for every Escherichia species and all phylogroups of E. coli.

Names of our isolates and their reference strains are coloured according to their species/phylogroup. Names of reference strains for clades not found in our samples are coloured black. Reference strains are named with their species and strain name, with phylogroups given in parentheses. Salmonella bongori and Salmonella. enterica ser. Typhi are included as outgroups.

Figure 3. Anvi’o representation of the Escherichia pangenome showing 81 Escherichia isolates recovered from chicken faeces and 33 publicly available reference Escherichia isolates from poultry species.

Each layer represents a single genome, with black colouring signifying the presence of a gene cluster. Gene clusters are organised according to their distribution across the genome, with co-occurring genes shown closer together. The heatmap at the top-right of the image represents average nucleotide identity across all included genomes, with darker red colours indicating a higher percentage of average nucleotide identity. Assigned phylogroup is shown by the colour bar on the right of the image. Singleton gene clusters (present only in one genome; n = 3,641) are highlighted in blue while core gene clusters (present in all genomes; n = 2,449) are shown in green.

Figure 4. Summary of the plasmid replicon contents of 81 Escherichia isolates.

(A) Number of isolates containing 0–7 distinct plasmid replicons from the PlasmidFinder database. (B) Frequency of each plasmid replicon identified from the PlasmidFinder database. Some isolates contained >1 of the same replicon type (see main text).

Figure 5. Presence (solid squares) and absence (open squares) of predicted antibiotic resistance genes, virulence-associated genes and plasmid replicons in the final 81 isolates.

Isolates are arranged according to the phylogenetic tree shown in Fig. 2 with their names coloured according to phylogenetic group. Ew, E. whittamii; Em, E. marmotae; Er, E. ruysiae; Ef, E. fergusonii; A–F, E. coli phylogroups. Virulence-associated genes shown here are from the custom panel of APEC-associated genes. Members of the panel not represented here were not present in any isolates. Plasmid replicon and predicted antibiotic resistance gene data have been condensed to show major classes. Therefore, presence indicates ≥1 members of that class were detected. Full gene and plasmid detection data are available in Tables S5–S7.