The intrinsic macrolide resistome of Escherichia coli

Abstract

Intrinsic resistance to macrolides in Gram-negative bacteria is primarily attributed to the low permeability of the outer membrane, though the underlying genetic and molecular mechanisms remain to be fully elucidated. Here, we used transposon directed insertion-site sequencing (TraDIS) to identify chromosomal non-essential genes involved in Escherichia coli intrinsic resistance to a macrolide antibiotic, tilmicosin. We constructed two highly saturated transposon mutant libraries of >290,000 and >390,000 unique Tn5 insertions in a clinical enterotoxigenic strain (ETEC5621) and in a laboratory strain (K-12 MG1655), respectively. TraDIS analysis identified genes required for growth of ETEC5621 and MG1655 under 1/8 MIC (n = 15 and 16, respectively) and 1/4 MIC (n = 38 and 32, respectively) of tilmicosin. For both strains, 23 genes related to lipopolysaccharide biosynthesis, outer membrane assembly, the Tol-Pal system, efflux pump, and peptidoglycan metabolism were enriched in the presence of the antibiotic. Individual deletion of genes (n = 10) in the wild-type strains led to a 64- to 2-fold reduction in MICs of tilmicosin, erythromycin, and azithromycin, validating the results of the TraDIS analysis. Notably, deletion of surA or waaG, which impairs the outer membrane, led to the most significant decreases in MICs of all three macrolides in ETEC5621. Our findings contribute to a genome-wide understanding of intrinsic macrolide resistance in E. coli, shedding new light on the potential role of the peptidoglycan layer. They also provide an in vitro proof of concept that E. coli can be sensitized to macrolides by targeting proteins maintaining the outer membrane such as SurA and WaaG.

Keywords: E. coli; TraDIS; intrinsic resistance; macrolide; tilmicosin.

Fig 1

Overview of genes required for growth of E. coli MG1655 and ETEC5621 after exposure to tilmicosin (TIL) at different concentrations. (A) UpSet plot showing intersections between strains and treatment combinations is displayed in a matrix layout, and each column corresponds to a specific section in a traditional Venn diagram. The histogram shows the number of unique and shared genes at each circle or intersection, respectively. Connected circles in the matrix indicate shared genes, and unconnected circles represent unique genes. Core genes of the intrinsic macrolide resistome are illustrated in red. (B) Main functions of the genes in the core resistome.

Fig 2

Growth curves of the wild-type (WT), mutant, and complementation strains of surA (A) and waaG (B) in ETEC5621. Optical density at 600 nm (OD₆₀₀) was measured over a 24-h growth period in Luria-Bertani (LB) broth control (CON) and LB broth supplemented with 8-µg/mL of tilmicosin (TIL). Data are the mean ± standard deviation of triplicate experiments.

Fig 3

Simultaneous time-lapse microscopy of ETEC5621 wild-type (WT) strain and mutant derivatives ΔsurA, ΔwaaG, Δprc, and ΔnlpI, grown on Luria-Bertani media supplemented with 1.2% agarose at 37°C. Three positions were imaged per strain, and phase-contrast images were acquired every 5 minutes for 24 h. Here 0, 2, 4 and 24 h are depicted for each strain.