Diversity and phage sensitivity to phages of porcine enterotoxigenic Escherichia coli

Abstract

Enterotoxigenic Escherichia coli (ETEC) is a diverse and poorly characterized E. coli pathotype that causes diarrhea in humans and animals. Phages have been proposed for the veterinary biocontrol of ETEC, but effective solutions require understanding of porcine ETEC diversity that affects phage infection. Here, we sequenced and analyzed the genomes of the PHAGEBio ETEC collection, gathering 79 diverse ETEC strains isolated from European pigs with post-weaning diarrhea (PWD). We identified the virulence factors characterizing the pathotype and several antibiotic resistance genes on plasmids, while phage resistance genes and other virulence factors were mostly chromosome encoded. We experienced that ETEC strains were highly resistant to Enterobacteriaceae phage infection. It was only by enrichment of numerous diverse samples with different media and conditions, using the 41 ETEC strains of our collection as hosts, that we could isolate two lytic phages that could infect a large part of our diverse ETEC collection: vB_EcoP_ETEP21B and vB_EcoS_ETEP102. Based on genome and host range analyses, we discussed the infection strategies of the two phages and identified components of lipopolysaccharides ( LPS) as receptors for the two phages. Our detailed computational structural analysis highlights several loops and pockets in the tail fibers that may allow recognition and binding of ETEC strains, also in the presence of O-antigens. Despite the importance of receptor recognition, the diversity of the ETEC strains remains a significant challenge for isolating ETEC phages and developing sustainable phage-based products to address ETEC-induced PWD.IMPORTANCEEnterotoxigenic Escherichia coli (ETEC)-induced post-weaning diarrhea is a severe disease in piglets that leads to weight loss and potentially death, with high economic and animal welfare costs worldwide. Phage-based approaches have been proposed, but available data are insufficient to ensure efficacy. Genome analysis of an extensive collection of ETEC strains revealed that phage defense mechanisms were mostly chromosome encoded, suggesting a lower chance of spread and selection by phage exposure. The difficulty in isolating lytic phages and the molecular and structural analyses of two ETEC phages point toward a multifactorial resistance of ETEC to phage infection and the importance of extensive phage screenings specifically against clinically relevant strains. The PHAGEBio ETEC collection and these two phages are valuable tools for the scientific community to expand our knowledge on the most studied, but still enigmatic, bacterial species-E. coli.

Keywords: enterotoxigenic E. coli; phage defense mechanisms; phage receptors; phages.

Fig 1

Phylogenetic tree built with the core genomes of the ETEC collection. Colors on the clades indicate the phylogroups, determined by submitting the ECOR genome sequences to EzClermont (https://ezclermont.hutton.ac.uk/). From left to right, we report the serotype (O- and H-type), year (green), country (blue), fimbriae (orange), and toxins (red).

Fig 2

Phage defense mechanisms predicted in the ETEC collection. Colors indicate the predicted position of the gene: yellow on chromosomes, orange on plasmids, blue on prophages. The histograms at the bottom represent the count of ETEC strains where each defense system was detected.

Fig 3

Host range analysis of the ETEC strains. For each strain, the phylogroup, the O-type, the virulence profile, and the infection by phages ETEP21B, ETEP102, and T7 are indicated.

Fig 4

Two diverse phages infect the ETEC strains. (A) Intergenomic similarity between ETEP21B, ETEP102, and other 158 genomes (their closest relatives, one genome for each of the genera within the same family, and one genome for each of the other families as listed in ICTV). Close up of (B) ETEP21B and (C) ETEP102 clusters. For each graph: on the right side, darker colors indicate higher intergenomic similarity and the numbers represent the similarity values for each genome pair; on the left side, for each genome pair are the aligned fraction genome versus other genomes in the same row, in the same column and the genome length ratio. Transmission electron micrographs of (D) phage ETEP21B and (E) ETEP102. The bar represents 100 nm, as indicated.

Fig 5

Genomic organization of phages ETEP21B (A) and ETEP102 (B). Genes are indicated as arrows: capsid morphogenesis in blue, tail morphogenesis in light-blue, DNA packaging in green, lysis-associated genes in pink, genes involved in DNA manipulation in red, morons and homing endonucleases in purple, recombination in yellow, and hypothetical in gray. Putative promoters are indicated as orange lines and terminators as brown lines.

Fig 6

(A) AlphaFold2-multimer model of ETEP21B-TF, colored by pLDDT confidence scores. (B) Putative second receptor-binding domain in T7-TF and ETEP21B-TF colored by ligand interface propensity and (C) ion interaction propensity. (D) Tips of T7-TF and ETEP21B-TF colored by ligand interface propensity.

Fig 7

(A) AlphaFold2 model for the ETEP102 tail fiber colored by pLDDT confidence scores. (B) Comparison of ETEP102 fiber depolymerase with the α-1,3-glucanase of Paenibacillus sp. (PDB-ID 6k0n) and CBA120 TSP2 (PDB-ID 5W6P), colored by ligand interface propensity. (C) Alignment of ETEP102-TF (right) with CBA120 TSP2 (left), showing that residues with high ligand interface propensity in ETEP102 are located closer to the N-terminus compared with the residues known to bind O-157 in TSP2 (in gray). The root mean square deviation between the two structures is 1.12 angstroms, considering 90 pruned atom pairs.