The population genomics of increased virulence and antibiotic resistance in human commensal Escherichia coli over 30 years in France

Abstract

Escherichia coli is a commensal species of the lower intestine, but also a major pathogen causing intestinal and extra-intestinal infections. Most studies on genomic evolution of E. coli used isolates from infections, and/or focused on antibiotic resistance, but neglected the evolution of virulence. Here instead, we whole-genome sequenced a collection of 436 E. coli isolated from fecal samples of healthy adult volunteers in France between 1980 and 2010. These isolates were distributed among 159 sequence types (STs), the five most frequent being ST10 (15.6%), ST73 (5.5%) and ST95 (4.8%), ST69 (3.7%) and ST59 (3.7%), and 230 O:H serotypes. ST and serotype diversity increased over time. Comparison with 912 E. coli bacteremia isolates from similar region and time showed a greater diversity in commensal isolates. The O1, O2, O6 and O25-groups used in bioconjugate O-antigen vaccine were found in only 63% of the four main STs associated with a high risk of bacteremia (ST69, ST73, ST95 and ST131). In commensals, STs associated with a high risk of bacteremia increased in frequency. Both extra-intestinal virulence-associated genes and resistance to antibiotics increased in frequency. Evolution of virulence genes was driven by both clonal expansion of STs with more virulence genes, and increases in frequency within STs, whereas the evolution of resistance was dominated by increases in frequency within STs. This study provides a unique picture of the phylogenomic evolution of E. coli in its human commensal habitat over a 30-year period and suggests that the efficacy of O-antigen vaccines would be threatened by serotype replacement.

The bacterial genetic determinants ofEscherichia colicapacity to cause bloodstream infections in humans

“…In fact, the most frequent co-morbidity in the BSI collection is immunosuppression, which was an exclusion criterion for the commensal collection. Comorbidities are associated with BSI (5,18,22). It is possible that co-morbidities act as a confounder in our study, if they both increase the probability of BSI and influence the E. coli strains carried by individuals.…”

“…Evolution of E. coli pathogenicity would have important public health implications, given that E. coli BSI are a major cause of morbidity and death in Western countries. To investigate temporal trends in pathogenicity, we computed the pathogenicity score with the machine learning model (used to predict the commensal vs. BSI status of strains), in a dataset of commensals from 1980 to 2010 in France (18). We found that the proportion of commensal E. coli isolates predicted to be pathogenic isolates with our trained model increased over time, from 23% in the collection from 1980 to 46% in the collection from 2010 (Figure S4).…”

“…In the present work, we took advantage of two recently published collections of BSI (1) and commensal (18) strains gathered between 2000 and 2017 in France and with their genome sequenced. We compared BSI and commensal strain genomes at three levels: phylogenomic composition, virulence and resistance gene content, and lastly unitig content in a GWAS.…”

“…So far, no studies investigated the bacterial genetic determinants of the capacity to cause an infection by comparing large numbers of whole genome sequences of bacteria sampled from the gut (commensals) vs. sampled from infections. This may be explained in part by the small number of large commensal strain collections (18). Another difficulty is that host factors, such as age or co-morbidities, are important determinants of infection (19), and must be adjusted for as much as possible when comparing strains sampled in the two contexts.…”

The E phylogroup of Escherichia coli is highly diverse and mimics the whole E. coli species population structure

“…Such a structure is also largely observed at the STc level, as for the STc95 (Gordon et al ., 2017). The time scale for these three taxonomic entities (species, phylogroup, ST) ranges from tens of millions of years (species) (Ochman and Selander, 1984) to hundreds of thousands of years (phylogroups) (Pupo et al ., 2000) and tens to hundreds of years (STs) (Stoesser et al ., 2016; Kallonen et al ., 2017; Franz et al ., 2019; Marin et al ., 2021). These data argue for a fractal genetic structure of the E. coli population from the species (Touchon et al ., 2009) to the STc (Gordon et al ., 2017) and suggests that similar evolutionary forces are acting at different scales and times.…”