Genomic landscape of extended-spectrum β-lactamase resistance in Escherichia coli from an urban African setting

Abstract

Objectives: Efforts to treat Escherichia coli infections are increasingly being compromised by the rapid, global spread of antimicrobial resistance (AMR). Whilst AMR in E. coli has been extensively investigated in resource-rich settings, in sub-Saharan Africa molecular patterns of AMR are not well described. In this study, we have begun to explore the population structure and molecular determinants of AMR amongst E. coli isolates from Malawi.

Methods: Ninety-four E. coli isolates from patients admitted to Queen's Hospital, Malawi, were whole-genome sequenced. The isolates were selected on the basis of diversity of phenotypic resistance profiles and clinical source of isolation (blood, CSF and rectal swab). Sequence data were analysed using comparative genomics and phylogenetics.

Results: Our results revealed the presence of five clades, which were strongly associated with E. coli phylogroups A, B1, B2, D and F. We identified 43 multilocus STs, of which ST131 (14.9%) and ST12 (9.6%) were the most common. We identified 25 AMR genes. The most common ESBL gene was bla CTX-M-15 and it was present in all five phylogroups and 11 STs, and most commonly detected in ST391 (4/4 isolates), ST648 (3/3 isolates) and ST131 [3/14 (21.4%) isolates].

Conclusions: This study has revealed a high diversity of lineages associated with AMR, including ESBL and fluoroquinolone resistance, in Malawi. The data highlight the value of longitudinal bacteraemia surveillance coupled with detailed molecular epidemiology in all settings, including low-income settings, in describing the global epidemiology of ESBL resistance.

Figure 1.

Genomic population structure and relatedness of E. coli isolates collected from Blantyre, Malawi. (a) Circular ML core-genome phylogenetic tree of Malawian E. coli isolates rooted at mid-point of the longest branch separating the two most divergent isolates. The inner ring designates identified SCs by hierBAPS and the outer ring designates phylogroups identified by in silico PCR. (b) Distribution of pairwise SNP differences in each of the five clades to demonstrate the variations in sequence diversity in each clade. (c) An ML phylogenetic tree, constructed from recombination-free SNP alignment of Malawian ST131 isolates (red branches) in the context of previously published global ST131 isolates (black branches). This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.

Figure 2.

Distribution of AMR phenotype profiles, acquired AMR genes and plasmid incompatibility groups across the phylogeny of Malawian E. coli isolates. On the left is the ML core-genome phylogeny of the E. coli isolates from Malawi. The first two columns at the termini of the phylogeny represent the clinical source of isolation and STs for each isolate, respectively. Immediately following the two columns are six columns that represent the AMR phenotype profile. The next two panels following the AMR phenotype profiles are columns that represent presence and absence of AMR genes and plasmid replicons. This figure appears in colour in the online version of JAC and in black and white in the printed version of JAC.