Having a specific combination of good genes in a complex interaction with the phylogenetic background of the strain is the magic recipe to transform a commensal bacterium to a major killer. We provide here strong evidence that epistasis, the interaction between genes influencing a phenotype, is at play in the emergence of extra-intestinal virulence in Escherichia coli. We show that the more virulent strains are those having the good genes coding mainly for iron capture systems, in a specific combination, at the right place in the genome (chromosome or plasmid-borne) and in a specific phylogenetic lineage. These findings open new ways of research to identify the molecular support of such associations.

E. coli is a major commensal species from vertebrates but also the first bacterial pathogen responsible for infection in humans. Several genes are known to be involved in its virulence. These so-called virulence associated genes (VAGs) can encode for adhesins, invasins, toxins, protectins or iron capture systems. Some of these genes are found in strains responsible for intestinal infections (intestinal pathogenic E. coli or InPEC) whereas other are more frequently observed in isolates from extra-intestinal infections (extra-instestinal pathogenic E. coli or ExPEC) such as urinary-tract infections or bacteremia¹. However, several factors limit our ability to clearly define the role of each of these genetic determinants in the pathogenicity such as i) the importance of host factors in the outcome of the disease^2,3, ii) the frequent combination of several VAGs in a given strain, iii) the additional role of the genetic background of the strain⁴, iv) the genomic location of these determinants (chromosomal or plasmidic). In this context, combining genomic data with phenotypic data from well-controlled animal models is an excellent opportunity to disentangle all these factors through a genome-wide association study.

We first used this kind of approach to get insight into the emergence of virulence within a particular clonal complex of E. coli, the CC87, often considered as commensal in animals⁵. Our data confirmed the emergence of a virulent and resistant lineage as previously suggested from genomic data only⁶, while highlighting the pivotal role of an iron capture system, the high pathogenicity island (HPI) or yersinibactin. Moreover, almost all the virulent strains also carry VAGs on an extra-chromosomal support, named ColV plasmid, that appears to have little or no role in our mouse model on infection.

Then, we extended our analysis to the whole genus Escherichia, taking advantage of previously published data combining whole genome sequences and phenotypic data in the same mouse model of sepsis⁷. We identified a strong association between the genetic background of the strains (phylogroups), the prevalence of the VAGs and their genomic location. Indeed, typical ExPEC from phylogroups B2 and D frequently carry VAGs on their chromosome, whereas in non-typical ExPEC or commensal, their prevalence was intermediate or low and they were mainly located on plasmids. By considering the phenotypic data, we were able to show that the VAGs do not have the same weight, but that there is in fact a hierarchy. Overall, iron capture systems appear to be the main drivers of virulence. Among them, three iron capture systems are particularly pivotal in the intrinsic virulence of strains, the major one being the HPI followed by sitABCD and the aerobatin (iucABCD/iutA). It should be noted that the last two can be found both on chromosome and ColV plasmid, without this having any significant impact on their association with virulence. A fourth, the salmochelin encoded by iroBCDEN and also found on both locations, does not play as important a role. Another striking feature of these VAGs is also their cumulative effect and the need for combinations of them to reach full virulence in the mouse model. Again, the HPI appears as a key factor required but not sufficient to express full virulence.

We then turned to another dataset consisting of more than 2,302 complete E. coli genomes to decipher these combinations of VAGs while taking into account their location based on high-quality genomic data. We were able to identify specific co-occurrence patterns for the main clones (identified by their sequence type, ST) responsible for bacteremia⁸. Some of them (e.g. ST131, ST73) achieve almost complete fixation of VAGs on their chromosome, in line with their ExPEC lifestyle⁹ and suggesting a niche adaptation. Conversely, others (e.g. ST10) only rarely carry VAGs and mainly on plasmids, suggesting transient pathogenicity and a versatile behaviour as usually observed in this species. Moreover, the different patterns of co-occurrences/location/genetic background also point to epistatic interactions with some cases of antagonism according to lineage/VAGs.

Overall, we found that multiple VAG associations are involved in E. coli extra-intestinal virulence. Iron capture systems are a major factor in virulence, with both a cumulative effect and a degree of hierarchy in their role. Depending on the phylogenetic lineage, the prevalence, the location and the combination of these determinants vary, supporting the role of epistasis in the emergence of virulence.

References

1. Denamur, E., Clermont, O., Bonacorsi, S. & Gordon, D. The population genetics of pathogenic Escherichia coli. Nat. Rev. Microbiol. 19, 37–54 (2021).
2. Lefort, A. et al. Host factors and portal of entry outweigh bacterial determinants to predict the severity of Escherichia coli bacteremia. J. Clin. Microbiol. 49, 777–783 (2011).
3. de Lastours, V. et al. Mortality in Escherichia coli bloodstream infections: antibiotic resistance still does not make it. J. Antimicrob. Chemother. 75, 2334–2343 (2020).
4. Picard, B. et al. The link between phylogeny and virulence in Escherichia coli extraintestinal infection. Infect. Immun. 67, 546–553 (1999).
5. Skurnik, D. et al. Emergence of antimicrobial-resistant Escherichia coli of animal origin spreading in humans. Mol. Biol. Evol. 33, 898–914 (2016).
6. Reid, C. J. et al. A role for ColV plasmids in the evolution of pathogenic Escherichia coli ST58. Nat. Commun. 13, 683 (2022).
7. Galardini, M. et al. Major role of iron uptake systems in the intrinsic extra-intestinal virulence of the genus Escherichia revealed by a genome-wide association study. PLoS Genet. 16, e1009065 (2020).
8. Royer, G. et al. Phylogroup stability contrasts with high within sequence type complex dynamics of Escherichia coli bloodstream infection isolates over a 12-year period. Genome Med. 13, 77 (2021).
9. Burgaya, J. et al. The bacterial genetic determinants of Escherichia coli capacity to cause bloodstream infections in humans. PLoS Genet. Accepted