Novel Insights into Selection for Antibiotic Resistance in Complex Microbial Communities

Abstract

Recent research has demonstrated that selection for antibiotic resistance occurs at very low antibiotic concentrations in single-species experiments, but the relevance of these findings when species are embedded in complex microbial communities is unclear. We show that the strength of selection for naturally occurring resistance alleles in a complex community remains constant from low subinhibitory to above clinically relevant concentrations. Selection increases with antibiotic concentration before reaching a plateau where selection remains constant over a 2-order-magnitude concentration range. This is likely to be due to cross protection of the susceptible bacteria in the community following rapid extracellular antibiotic degradation by the resistant population, shown experimentally through a combination of chemical quantification and bacterial growth experiments. Metagenome and 16S rRNA analyses of sewage-derived bacterial communities evolved under cefotaxime exposure show preferential enrichment for bla_CTX-M genes over all other beta-lactamase genes, as well as positive selection and co-selection for antibiotic resistant, opportunistic pathogens. These findings have far-reaching implications for our understanding of the evolution of antibiotic resistance, by challenging the long-standing assumption that selection occurs in a dose-dependent manner.IMPORTANCE Antibiotic resistance is one of the greatest global issues facing society. Still, comparatively little is known about selection for resistance at very low antibiotic concentrations. We show that the strength of selection for clinically important resistance genes within a complex bacterial community can remain constant across a large antibiotic concentration range (wide selective space). Therefore, largely understudied ecological compartments could be just as important as clinical environments for selection of antibiotic resistance.

Keywords: antibiotic resistance; evolution; metagenomics; microbial ecology.

FIG 1

Heat map showing the relative abundance of detected species using Bray-Curtis distance measurements for treatment (x axis) and species (y axis) for each cefotaxime treatment. “C0,” “C4,” “C6,” and “C8” correspond to 0, 125, and 500 μg/liter and 2 mg/liter cefotaxime, respectively. The number after the concentration denotes the biological replicate number (1 to 5), chosen randomly for sequencing at day 8 of the experiment.

FIG 2

Heat map showing average (n = 3) detected beta-lactam resistance gene subtype relative abundance (resistance gene number normalized with 16S rRNA copy number), following 8 days of culture with cefotaxime. Only genes detected with the ARGs-OAP pipeline are shown.

FIG 3

Selection coefficients (n = 5) for each cefotaxime concentration, which equal the natural log of resistance gene prevalence (bla_CTX-M gene/16S rRNA copy number) at day 0 and day 8. Circles, selection coefficients from low-concentration experiment; squares, selection coefficients from high-concentration experiment. Selection coefficients of >0 indicate positive selection.

FIG 4

Single biological replicate and duplicate chemical replicate quantification of cefotaxime (measured cefotaxime concentration µg/liter) at 0 and 6 h and then every 3 h for 24 h at different starting cefotaxime concentrations (micrograms/liter) in the presence of the complex bacterial community and in sterile culture.

FIG 5

Average (n = 4) optical density (600 nm) over time of susceptible E. coli strain J53 grown in 2 mg/liter cefotaxime (clinical breakpoint concentration for Enterobacteriaceae) with beta-lactamase-containing supernatant (NCTC strain 13451) or beta-lactamase-free supernatant (strain J53).