Molecular mechanisms of re-emerging chloramphenicol susceptibility in extended-spectrum beta-lactamase-producing Enterobacterales

Abstract

Infections with Enterobacterales (E) are increasingly difficult to treat due to antimicrobial resistance. After ceftriaxone replaced chloramphenicol (CHL) as empiric therapy for suspected sepsis in Malawi in 2004, extended-spectrum beta-lactamase (ESBL)-E rapidly emerged. Concurrently, resistance to CHL in Escherichia coli and Klebsiella spp. decreased, raising the possibility of CHL re-introduction. However, many phenotypically susceptible isolates still carry CHL acetyltransferase (cat) genes. To understand the molecular mechanisms and stability of this re-emerging CHL susceptibility we use a combination of genomics, phenotypic susceptibility assays, experimental evolution, and functional assays for CAT activity. Here, we show that of 840 Malawian E. coli and Klebsiella spp. isolates, 31% have discordant CHL susceptibility genotype-phenotype, and we select a subset of 42 isolates for in-depth analysis. Stable degradation of cat genes by insertion sequences leads to re-emergence of CHL susceptibility. Our study suggests that CHL could be reintroduced as a reserve agent for critically ill patients with ESBL-E infections in Malawi and similar settings and highlights the ongoing challenges in inferring antimicrobial resistance from sequence data.

Fig. 1. Characteristics of isolates.

a Phenotypic CHL susceptibility of E. coli and KpSC isolates based on disc diffusion. b Number of CHL susceptible isolates carrying cat, floR or cmlA genes, both (cat + floR or cmlA) or no CHL resistance genes. c Number of CHL resistance genes present in the 840 Malawian isolates. Source data are provided as a Source Data file.

Fig. 2. CAT enzyme activity.

CAT enzyme activity was measured using the rCAT assay. The difference ((signal + CHL) − (signal − CHL) for a single isolate (n = 2–6)) in absorbance at 405 nm is given for each of the 42 isolates. The colour indicates if the isolate is phenotypically resistant (R, blue) or susceptible (S, yellow) to CHL based on broth microdilution and the shape indicates E. coli (circle) or K. pneumoniae (triangle). The Y-axis is ordered according to the presence of cat genes. Source data are provided as a Source Data file.

Fig. 3. Cat gene degradation by insertion sequences.

a Schematic of IS26 truncation of catB3. b Alignment of four contigs from four different isolates containing either catB3 or catB3∆^443–633. c Alignment of two genomes containing catA1, with and without an IS5 insertion into the catA1 promoter.

Fig. 4. Co-occurrence network of selected AMR genes.

Heatmap displaying co-occurrence relationships between AMR genes as either positive (blue), random (grey) or negative (orange). These are probabilistic values based on the difference in expected and observed frequencies of co-occurrence between each pair of genes, these values were obtained by applying the probabilistic model from. Co-occurrence across select genes including cat genes, aac(6’)-Ib-cr, bla_OXA-1, bla_CTX-M-15, and bla_TEM-1. Source data are provided as a Source Data file.

Fig. 5. Cat gene occurrence and proportions per ST.

E. coli isolates from Malawi included in our study (top panel) and in a 10k genome collection of 100 randomly selected genomes of the top 100 most common STs from E. coli (bottom panel). Source data are provided as a Source Data file.