Abstract :
[en] O80:H2 is an emerging Shigatoxigenic Escherichia coli (STEC) serotype since 2010 in humans in Western Europe. STEC O80:H2 was previously considered as a “minor” serotype, much less frequently associated with the haemolytic-uremic syndrome (HUS) than the classical "major" serotypes, such as O26:H11 and O157:H7. The primary virulence properties of STEC O80:H2 include (i) the production of Shiga toxins (Stx) encoded by phage-borne stx genes (e.g., stx1a, stx2a, stx2c, and stx2d…), and (ii) the ability to form attaching and effacing (A/E) lesions via a Type3 Secretion System (T3SS) and the intimin adhesin encoded by genes located on a pathogenicity island called “Locus of Enterocyte Effacement” (LEE). Since these STEC produce the A/E lesion, they will be named AE-STEC from now on. The eae gene coding for the intimin adhesin of AE-STEC O80:H2 belongs to the specific eaeXI (eae) variant. However, unlike the major serotypes, STEC O80:H2 are distinct in their involvement not only in hemorrhagic colitis (HC) and HUS, but also in systemic infections, such as bacteremia and septicemia, due to additional invasiveness factors encoded by genes located on ColV pS88-like plasmids. Furthermore, not only AE-STEC O80:H2, but also enteropathogenic E. coli (EPEC) O80:H2, producing the A/E lesion but no Stx, have been isolated more frequently since 2009 from diarrheic, and occasionally septicemic, calves in Belgium, suggesting a possible animal reservoir. However, although asymptomatic cattle and other ruminants are indeed carriers of the major AE-STEC serotypes, the true reservoir of STEC O80:H2 remains unidentified, due to their limited or absent association with adult healthy cattle and with bovine-related foodstuffs.
The general purpose of this Ph.D. work was therefore to obtain more information on the structural and functional understanding of E. coli O80:H2 population by: (i) renewing attempts to isolate AE-STEC and/or EPEC from asymptomatic adult cattle at slaughterhouses and in farms (STUDIES 1 and 2); (ii) confirming the genome structure, the virulotypes and the phylogenesis of human and calf AE-STEC and EPEC O80:H2 populations (STUDY 3); and (iii) assessing the respective roles of different properties of AE-STEC and EPEC O80:H2 in an in vivo model in larvae of Galleria mellonella moth (STUDY 4).
STUDIES 1 and 2 focused on attempting to isolate AE-STEC or EPEC O80:H2 from healthy cattle using both non-specific (STUDY 1) and trait-based specific (STUDY 2: non-melibiose fermentation and potassium tellurite resistance) methods. Although some O80 PCR-positive fecal samples were recovered from healthy cattle at two slaughterhouses and from healthy cows in nine farms during STUDY 1, none of the isolates belonged to the serotype O80:H2. Whole genome sequencing confirmed that these E. coli O80:nonH2 belonged to serotypes O80:H6 and O80:H45, that are phylogenetically distant from serotype O80:H2, and were non-AE-STEC non-EPEC. In STUDY 2, all 52 human and calf AE-STEC and EPEC O80:H2 analysed did not harbour the mel operon and were non-melibiose fermentative (only calf isolates were phenotypically tested). Conversely, the resistance levels to potassium tellurite and the presence of the ter operon were different among the human and calf AE-STEC and EPEC. Most of the human and calf AE-STEC O80:H2 did not harbor the ter operon and had low Minimal Inhibitory Concentrations (MIC) to tellurite, in contrast to most calf EPEC (only calf isolates were phenotypically tested). Applying these two properties on faecal samples from one slaughterhouse, none of the isolates O80 belonged to the serotype O80:H2.
STUDY 3 aimed to clarify (i) the combination of virulence-associated genes (virulotypes) and the population phylogenetic structure by expanding the set of analyzed human and calf AE-STEC and EPEC O80:H2 from Belgium from 52 to 129 and (ii) the virulence gene localization by generating complete genomic sequences of two stx2f-positive calf AE-STEC O80:H2. Besides the detection of the stx2f gene in two calf AE-STEC in addition to the stx1a, stx2a and stx2d genes, the virulotypes of calf and human AE-STEC and EPEC were confirmed by the detection of (i) the eaeξ LEE-located gene coding for the intimin adhesin; (ii) different T3SS-encoding (non)-LEE-located genes (espA, espB, espF, tir/nleA, nleB, nleC); (iii) the pS88-located (cia, cvaA, etsC, iucC, hlyF, ompTp, iss, iroN, sitA) genes; (iv) several other plasmid- or chromosome-located virulence-associated genes (cia, imm, espP, ehxA, efa1...). The population structure was also confirmed with the presence of two lineages (L) in a Single Nucleotide Polymorphism (SNP) phylogenetic tree and with L1 further subdivided in four sub-lineages (SL), whom SL1.1 and SL1.2 grouped 96% of the human and calf AE-STEC and EPEC O80:H2. Moreover, two virulence-associated gene profiles differentiated the strains of SL1.1 and SL1.2 based on the presence / absence of genes or gene clusters located, in theory, on the pS88 plasmids (ets, iuc) or not (cma, iha).
Complete genomic sequences of the two calf stx2f AE-STEC (SES0057 belonging to SL1.1 and SES0108 belonging to SL1.2) was generated by long-read sequencing to clarify the genome structures of the LEE regions, of the Stx2f phages and of the pS88 plasmids, and to confirm the localization of the ets, iuc, cma, iha genes. At first, the LEE regions were identical. Two identical copies of the Stx2f were detected in either strain with one copy inserted within the ssrA tmRNA locus and the second one integrated into another prophage at the height of the thrW tRNA locus (prophage-in-prophage). The Stx2f phages of strain SES0108 were >99% identical to the Stx2f phages of two E. coli strains belonging to other serotypes while the Stx2f phages of strain SES0057 were different. The pS88 plasmid of strain SES0057 belonging to SL1.1 did not carry the ets and iuc genes that had not been detected in this strain, but surprisingly carried the cma gene, while the pS88 plasmid of strain SES0108 belonging to SL1.2 carried the ets and iuc genes, but not the cma gene that had not been detected in this strain. Finally, the iha gene detected only in strain SES0057 was located on the chromosome-integrated “Sakai-prophage like element-1” (SpLE-1).
STUDY 4 assessed the functional contribution of different factors to the virulence of AE-STEC and EPEC O80:H2 by testing them in the larvae of the moth Galleria mellonella. Three stx1a AE-STEC, three stx2d AE-STEC and two EPEC strains harboring different pS88 plasmids, one K12 DH10B laboratory E. coli strain transduced with the Stx2d phage, one K12 DH10B laboratory E. coli strain conjugated with a pS88 plasmid carrying the ets and iuc genes, and three control strains were tested in this in vivo model. The stx2d AE-STEC and Stx2d-transduced K12 DH10B strains had the statistically significant highest lethality of all, but the positive control strains suggesting that the Stx2d is a major virulence factor in this model. The virulence of the Stx1a AE-STEC was not significantly higher than the virulence of the EPEC strains, leaving open the question of the actual role of Stx1a in this model. Similarly, the contribution of the different pS88 plasmids to the virulence of the different AE-STEC strains was statistically not significant, while the contribution of the pS88 plasmid carrying the ets and iuc genes, but not of the pS88 plasmid not carrying these two genes, to the virulence of the EPEC and of the transconjugated K12 DH10B strains was slightly significant. These results suggest that the pS88 plasmids do not highly contribute to the virulence of AE-STEC and EPEC O80:H2 in the Galleria mellonella larva model.
In conclusion, STUDIES 1 and 2 about isolation and identification highlight that non-melibiose fermentation, but not resistance to tellurite could help to isolate AE-STEC and EPEC O80:H2 from different sources, especially when present in low numbers like probably in healthy cattle carriers. Therefore, several other methodologies, not only agar-dependent conventional or specific ones testing different antibiotics and antiseptics, but also agar-independent ones, should be developed and assessed in future surveys. In STUDY 3, the population structure of Belgian human and calf AE-STEC and EPEC O80:H2, including the phylogenetic classification (two lineages with one of them subdivided in four sub-lineages) and the virulotypes (presence of a LEE in all but one strain, of different stx genes in the STEC and of a pS88 plasmid), was confirmed when extending the number of strains studied from 52 to 129. The existence of two gene profiles according to the belonging to one of the sub-lineages based on the presence / absence of, in theory pS88-located (iuc and ets) or not (cma and iha) genes or gene clusters was also confirmed. Also the features and the genome structures of the LEE regions, Stx2f phages, and pS88 plasmids were clarified by generating and analyzing the whole genome sequences of two calf stx2f AE-STEC belonging to the sub-lineages with different gene profiles. Surprisingly, the cma gene was also found to be located on the pS88 plasmid in the positive strain. Lastly, STUDY 4 assessed the contribution of several virulence properties using larvae of the moth Galleria mellonella as an in vivo model: the Stx2d is the most virulent property, while the two different pS88 plasmids do not highly contribute to the virulence in this insect model. The contribution of Stx1a to virulence in this model, if any, was not possible to assess at this stage.
