Phylogenomic Analysis of Extraintestinal Pathogenic Escherichia coli Sequence Type 1193, an Emerging Multidrug-Resistant Clonal Group

Abstract

The fluoroquinolone-resistant sequence type 1193 (ST1193) of Escherichia coli, from the ST14 clonal complex (STc14) within phylogenetic group B2, has appeared recently as an important cause of extraintestinal disease in humans. Although this emerging lineage has been characterized to some extent using conventional methods, it has not been studied extensively at the genomic level. Here, we used whole-genome sequence analysis to compare 355 ST1193 isolates with 72 isolates from other STs within STc14. Using core genome phylogeny, the ST1193 isolates formed a tightly clustered clade with many genotypic similarities, unlike ST14 isolates. All ST1193 isolates possessed the same set of three chromosomal mutations conferring fluoroquinolone resistance, carried the fimH64 allele, and were lactose non-fermenting. Analysis revealed an evolutionary progression from K1 to K5 capsular types and acquisition of an F-type virulence plasmid, followed by changes in plasmid structure congruent with genome phylogeny. In contrast, the numerous identified antimicrobial resistance genes were distributed incongruently with the underlying phylogeny, suggesting frequent gain or loss of the corresponding resistance gene cassettes despite retention of the presumed carrier plasmids. Pangenome analysis revealed gains and losses of genetic loci occurring during the transition from ST14 to ST1193 and from the K1 to K5 capsular types. Using time-scaled phylogenetic analysis, we estimated that current ST1193 clades first emerged approximately 25 years ago. Overall, ST1193 appears to be a recently emerged clone in which both stepwise and mosaic evolution have contributed to epidemiologic success.

Keywords: ExPEC; ST1193; antimicrobial resistance; fluoroquinolone; plasmid.

FIG 1

Phylogenetic tree of ST1193 (n = 355) and non-ST1193 STc14 (n = 72) isolates based on single-nucleotide polymorphisms in nonrecombinant core genome regions. In the tree, bright green is ST1193 with IncF-:A1:B1 plasmids, red is ST1193 with IncF-:A1:B10 plasmids, pink is ST1193 with IncF-:A1:B20 plasmids, and black is non-ST1193, STc14. Blue-colored circles indicate the presence of specific capsular types, and purple-colored squares indicate presence of F-plasmid alleles. Isolate names colored green indicate the ST1193 isolates sequenced in this study.

FIG 2

Phylogenetic distribution of accessory traits in representative ST1193 and non-ST1193 STc14 isolates. Squares, colored by trait category, represent the presence of a trait examined. Mutations refer to mutations conferring fluoroquinolone resistance. Year (green to red, older to recent) and ST are categorically colored for comparison. Plasmid mapping refers to the presence of a region (>95% similarity and >80% length) relative to plasmid p1ESCUM.

FIG 3

Mapping of sequencing reads from ST1193 and non-ST1193 STc14 isolates against plasmid pNMEC-O75A in the context of F-plasmid replicon. In the sequencing maps, black peaks indicate regions with exact matches, colored peaks indicate matching regions with nucleotide variation, and white regions indicate areas with no mapping of reads.

FIG 4

Time-scaled phylogenetic analysis of 111 ST1193 isolates collected in the U.S. between 2008 and 2018. Red node bars indicate the 95% highest posterior density interval of the node position. Purple circles indicate nodes with at least 70% posterior support. Brown bars at the right indicate the presence of capsular types K1 and K5, and purple bars at right indicate presence of F plasmid alleles.

FIG 5

Heatmap presence or absence of non-core-predicted proteins across STc14 isolates sequenced in this study. Red, presence of predicted protein presence; cream, absence of predicted protein. Heatmap was generated using Ward two-way hierarchical clustering with the Euclidean distance method.

FIG 6

Enterobase GrapeTree of 484 STc14 isolates, generated using core genome multilocus sequence typing (cgMLST). Using the Achtman 7-gene MLST scheme in Enterobase, a search for ST14 with a mismatch setting of 1 allele resulted in a database of 483 STc14 strains. cgMLST was used in the experimental data setting to generate GrapeTree (minimal spanning tree) images. The upper image is colored according to ST, and the lower image is colored according to predicted serogroup. In the keys, the numbers in brackets represent number of isolates.

FIG 7

Postulated key events leading to the currently circulating ST1193 fluoroquinolone-resistant clonal group. RepFIB alleles B1, B10, and B20 are depicted as boxes at the lower right.