Gallium-mediated siderophore quenching as an evolutionarily robust antibacterial treatment

Abstract

Background and objectives: Conventional antibiotics select strongly for resistance and are consequently losing efficacy worldwide. Extracellular quenching of shared virulence factors could represent a more promising strategy because (i) it reduces the available routes to resistance (as extracellular action precludes any mutations blocking a drug's entry into cells or hastening its exit) and (ii) it weakens selection for resistance, as fitness benefits to emergent mutants are diluted across all cells in a cooperative collective. Here, we tested this hypothesis empirically.

Methodology: We used gallium to quench the iron-scavenging siderophores secreted and shared among pathogenic Pseudomonas aeruginosa bacteria, and quantitatively monitored its effects on growth in vitro. We assayed virulence in acute infections of caterpillar hosts (Galleria mellonella), and tracked resistance emergence over time using experimental evolution.

Results: Gallium strongly inhibited bacterial growth in vitro, primarily via its siderophore quenching activity. Moreover, bacterial siderophore production peaked at intermediate gallium concentrations, indicating additional metabolic costs in this range. In vivo, gallium attenuated virulence and growth-even more so than in infections with siderophore-deficient strains. Crucially, while resistance soon evolved against conventional antibiotic treatments, gallium treatments retained their efficacy over time.

Conclusions: Extracellular quenching of bacterial public goods could offer an effective and evolutionarily robust control strategy.

Keywords: Pseudomonas; antivirulence therapy; experimental evolution; public good quenching; resistance.

Figure 1.

Gallium affects P. aeruginosa’s in vitro growth and siderophore production. (A) Gallium suppresses growth particularly when pyoverdine is present, as shown here by comparing conditions with and without its production. Symbols and bars indicate means and 95% CIs of integrals of spline curves fitted through 24 h growth trajectories (OD at 600 nm) of 12 replicate cultures. (B) Pyoverdine, assayed using complementary approaches, is in each case upregulated at intermediate gallium concentrations. Symbols and error bars represent means and SEs of five replicates. Measures of pyoverdine from supernatant (filled circles) or pvdA expression from cell fractions (open circles) are in each case scaled by cell density (OD at 600 nm).

Figure 2.

Gallium attenuates P. aeruginosa virulence and growth in G. mellonella larvae. (A–C) Virulence across treatments, as Kaplan–Meier (stepped lines) and Weibull (smoothed lines) survival curves; proportion surviving (with 95% binomial CIs); and time-to-death (means and 95% CIs). We estimate that inocula with ‘LOW’ (2.5–50 μM), ‘MED’ (500 μM) or ‘HIGH’ (2500 μM) concentrations of Ga(NO₃)₃ gave in-host concentrations of ∼0.05 to ∼50 μM (see ‘Methodology’ section). (D) Bacterial density in vivo (GFP signal in host homogenate; means and 95% CIs from ∼24 larvae) corrected against saline-injected controls and scaled relative to PAO1 at 13 h. (E) Mean and 95% CIs of bacterial growth integrals derived from bootstrap replicate time series (24 replicate splines) from (D)

Figure 3.

Evolutionary potential for resistance against gallium treatment. (A–D) Over the course of experimental evolution, daily growth integrals for cultures treated with various antibiotics rose significantly, while the growth of gallium treated cultures did not. (E) Slope coefficients for linear fits through data in (A–D), expressed as % of growth of control at Day 1. In all cases, symbols and error bars show means and 95% CIs of six replicate cultures

Figure 4.

Resistance-related phenotypic changes following experimental evolution under gallium treatment. Pyoverdine (A) and pyocyanin (B) production under standardized test conditions (dark bars = LB medium, light bars = CAA medium) of ancestral PAO1, knock-out strains (i.e. negative controls), control lines (evolved without gallium) and gallium-selected lines. Pyoverdine measures are scaled to that of PAO1 in CAA, whereas pyocyanin is scaled to that of PAO1 in LB. Asterisks indicate cases where Ga-selected lines were significantly different from their ancestor and unexposed control lines. Error bars give 95% CIs of 3–6 replicates