The chemical ecology of coumarins and phenazines affects iron acquisition by pseudomonads

Abstract

Secondary metabolites are important facilitators of plant-microbe interactions in the rhizosphere, contributing to communication, competition, and nutrient acquisition. However, at first glance, the rhizosphere seems full of metabolites with overlapping functions, and we have a limited understanding of basic principles governing metabolite use. Increasing access to the essential nutrient iron is one important, but seemingly redundant role performed by both plant and microbial Redox-Active Metabolites (RAMs). We used coumarins, RAMs made by the model plant Arabidopsis thaliana, and phenazines, RAMs made by soil-dwelling pseudomonads, to ask whether plant and microbial RAMs might each have distinct functions under different environmental conditions. We show that variations in oxygen and pH lead to predictable differences in the capacity of coumarins vs phenazines to increase the growth of iron-limited pseudomonads and that these effects depend on whether pseudomonads are grown on glucose, succinate, or pyruvate: carbon sources commonly found in root exudates. Our results are explained by the chemical reactivities of these metabolites and the redox state of phenazines as altered by microbial metabolism. This work shows that variations in the chemical microenvironment can profoundly affect secondary metabolite function and suggests plants may tune the utility of microbial secondary metabolites by altering the carbon released in root exudates. Together, these findings suggest that RAM diversity may be less overwhelming when viewed through a chemical ecological lens: Distinct molecules can be expected to be more or less important to certain ecosystem functions, such as iron acquisition, depending on the local chemical microenvironments in which they reside.

Keywords: coumarin; phenazine; pseudomonas; redox; secondary metabolite.

Fig. 1.

The chemical ecology of rhizosphere redox-active metabolites (RAMs) may help resolve their seemingly redundant functions. (A) Plant coumarins (green) and bacterial phenazines (orange) are both RAMs that aid in iron solubilization, yet it is unknown whether changes in environmental chemistry may favor one type of metabolite over another. One notable difference between the metabolites is that phenazines are more reactive with oxygen than coumarins. (B) We can predict distinct chemical ecological niches where coumarins and phenazines may be more effective at promoting iron acquisition over a regime defined by variable oxygen, pH, and carbon exudate concentrations. Phenazines (orange) may be most useful in environments where they can be easily reduced by microbes: under low pH, low O₂, and in the presence of reduced carbon sources. Coumarins (green) are not as reactive with oxygen and do not require microbial reduction to liberate iron, potentially making them useful across a wide range of environmental conditions. (C_ox): carbon oxidation state.

Fig. 2.

Chemical properties of coumarins and the phenazine PCA. (A) Voltammograms showing reversible redox activity of daphnetin, fraxetin, and PCA at pH 5.5, 6.5, and 7.5. Experiments were conducted with 1 mM metabolite under ambient oxygen (coumarins) or N₂ (PCA). The CV profile of scopoletin is complex (37), see SI Appendix, Fig. S1. Coumarin CV experiments used a gold working electrode, PCA CV used a glassy carbon working electrode (Materials and Methods). Note large scale difference in potentials between coumarins and PCA. (B) Iron reduction by different metabolites in the presence (black) or absence (blue) of oxygen. Experiments were conducted at pH 7.5 with 40 µM Fe(III) as FeCl₃ and 20 µM of each metabolite, Fe(II) was detected with the ferrozine assay and shows 2:1 metabolite:Fe(II) stoichiometry. See SI Appendix, Figs. S2 and S3 for further controls and experiments at pH 5.5. Duplicate assays are shown as two separate lines. (C) Structures of oxidized and reduced coumarins and PCA. Redox reactions involving scopoletin are complex and only one potential oxidation product is shown (37). Metabolites used in A and B are indicated in panel C.

Fig. 3.

The coumarins daphnetin and fraxetin stimulate growth of iron-limited pseudomonads. Effect of coumarins and the phenazine PCA on growth of P. synxantha (A) and P. aeruginosa (B). Under iron-limitation growth yields of both species were enhanced in the presence of daphnetin and fraxetin but not scopoletin, coumarin or PCA (A and B). Under nitrogen limitation, coumarin additions had mildly toxic effects on growth yields (A and B). All experiments used succinate as the carbon source and were conducted at pH 7.5 under ambient oxygen with fully air-equilibrated metabolites, which in the case of PCA leads to metabolite oxidation. The growth medium for Fe-limited experiments used hydrous ferric oxide (HFO) as an insoluble iron source and contained 16 mM N. The growth medium for nitrogen-limited experiments used 100 µM EDTA and 10 µM Fe as a soluble iron source and contained 2 mM N. Data shown are for biological duplicates ±SD. Some error bars may not be visible if they are smaller than the symbol. See also SI Appendix, Fig. S4 for additional Fe-limited experiments. Statistical analyses (one-way ANOVA, with Tukey post hoc correction) were performed on data from the final timepoint (Inset) and denote significant differences between the indicated treatment and the no addition treatment.

Fig. 4.

PCA stimulates growth of pseudomonads under mildly acidic hypoxic conditions. (A and B) Growth of aerobic cultures of P. synxantha (A) and P. aeruginosa (B) maintained under continuous shaking (black) or allowed to sit statically for 1 h (blue, arrow indicates time point at which a 1 h static incubation was conducted) with: no addition, 100 µM PCA, or 100 µM fraxetin. Data shown are biological duplicates ±SD. (C) Oxygen consumption during P. synxantha 1 h static incubations with PCA (conducted at time point indicated by arrow in A) at pH 5.5 and 7.5. Each line represents one biological duplicate (growth data shown in A). (D) P. synxantha short-term PCA reduction rates at pH 5.5 and 7.5. To separate the effects of oxygen and pH, experiments were conducted in an anoxic chamber and reflect rates obtained over 60 min. Data shown are from quadruplicate incubations ±SD. All experiments (A–D) were conducted under iron limitation with HFO as the iron source and succinate as the carbon source. Significant differences between shaking and static treatments ((A and B), two-way ANOVA) are shown for each time point. Difference between PCA reduction rates were calculated using a two-way ANOVA comparing the effects of carbon source and pH (Materials and Methods and SI Appendix, Fig. S6).

Fig. 5.

Carbon source tunes the effects of PCA on P. synxantha. Growth of P. synxantha with HFO as the iron source and glucose (A), succinate (B), or pyruvate (C) as the carbon source. Metabolites were added at a final concentration of 100 µM and experiments were conducted under ambient oxygen conditions. Slow shaking speeds allow microbial respiration to decrease oxygen, while faster speeds maintain higher oxygen tensions. When succinate and pyruvate are the carbon source PCA primarily stimulates growth yields at slow shaking speeds and at pH 5.5. However, growth on glucose allows PCA to enhance growth yields at high shaking speeds and pH 7.5. Three biological replicates are shown as individual lines. Statistical significance is shown in SI Appendix, Fig. S7. See also SI Appendix, Fig. S8 for replicate experiments showing consistency in broad trends. Total carbon was kept constant across treatments.