Uropathogenic Escherichia coli wield enterobactin-derived catabolites as siderophores

Abstract

Uropathogenic Escherichia coli (UPEC) secrete multiple siderophore types to scavenge extracellular iron(III) ions during clinical urinary tract infections, despite the metabolic costs of biosynthesis. Here, we find the siderophore enterobactin (Ent) and its related products to be prominent components of the iron-responsive extracellular metabolome of a model UPEC strain. Using defined Ent biosynthesis and import mutants, we identify lower molecular weight dimeric exometabolites as products of incomplete siderophore catabolism, rather than prematurely released biosynthetic intermediates. In E. coli, iron acquisition from iron(III)-Ent complexes requires intracellular esterases that hydrolyze the siderophore. Although UPEC are equipped to consume the products of completely hydrolyzed Ent, we find that Ent and its derivatives may be incompletely hydrolyzed to yield products with retained siderophore activity. These results are consistent with catabolic inefficiency as means to obtain more than one iron ion per siderophore molecule. This is compatible with an evolved UPEC strategy to maximize the nutritional returns from metabolic investments in siderophore biosynthesis.

Keywords: Escherichia coli; Gram-negative bacteria; enterobactin; exometabolome; iron; metabolomics; siderophore.

Figure 1

The enterobactin (Ent) biosynthetic pathway is a prominent contributor to the iron-responsive uropathogenic Escherichia coli UTI89 exometabolome. Sparse principal component analysis (sPCA) was performed to identify LC–MS exometabolome profiles that distinguish four groups of conditioned media: UTI89 grown in low and high iron (addition of 100 μM FeCl₃) media (wildtype and wildtype + Fe, respectively) and the Ent-null mutant UTI89ΔentB grown in low and high iron media (entB and entB + Fe, respectively). A, the score plot depicts each replicate LC–MS exometabolome (as a data point) as a function of principal components 1 and 2 (PC1, PC2), which are the first and second most influential modes of exometabolomic variation across all specimens. The combination of exometabolites comprising the PC1 axis separate the wildtype, low iron condition from the other conditions. B, LC–MS/MS chromatograms corresponding to the precursor–product ions from Ent (m/z 668.0) for each experimental group. Chromatograms are displayed in identical ion current unit scales. C, a PC1 loading plot displays the contribution of each LC–MS exometabolite feature to the PC1 value. The 13 most influential metabolites with greater abundance in wildtype UTI89 (lower PC1 value) are identified as red data points.

Figure 2

Exometabolites associated with the iron-responsive UTI89 exometabolome. A, chemical structures of the 10 enterobactin (Ent)-associated exometabolites identified by comparative metabolomic analysis, including Ent, monoglucosylated Ent (MGE), diglucosylated Ent (DGE), linear Ent (lin-Ent), linear monoglucosylated Ent (lin-MGE), linear diglucosylated Ent (lin-DGE), N-(2,3-dihydroxybenzoyl)serine dimer [(DHBS)₂], monoglucosylated N-(2,3-dihydroxybenzoyl)serine dimer [G₁-(DHBS)₂], diglucosylated N-(2,3-dihydroxybenzoyl)serine dimer [G₂-(DHBS)₂], and N-(2,3-dihydroxybenzoyl)serine monomer (DHBS). The positions of C-glucosylated DHBS units within linear polymers have not been definitively identified. B, heatmap showing Ent-associated exometabolite concentrations in media with iron supplementation or defined biosynthetic mutants of UTI89. Intensity represents concentration expressed as ratio of LC–MS/MS peak area to that of internal standard. Individual biological replicates are shown for each condition.

Figure 3

Outer membrane import differentially affects trimeric and nontrimeric enterobactin (Ent)-associated exometabolites in culture. Ent-associated exometabolite concentrations in media conditioned by UTI89 (wildtype), an import-deficient UTI89 mutant (ΔtonB), or coculture of Ent-null and import-deficient UTI89 mutants (ΔtonB + ΔentB). Y-axis is concentration expressed as ratio of LC–MS/MS peak area to that of internal standard. (A) Ent. (B) lin-Ent. (C) MGE. (D) lin-MGE. (E) DGE. (F) lin-DGE. (G) (DHBS)₂. (H) G₁-(DHBS)₂. (I) G₂-(DHBS)₂. (J) DHBS. Data are presented as mean ± SD with at least three biologically independent samples. Statistics were performed using one-way ANOVA with Dunnett’s multiple-comparison test with p ≤ 0.05 considered as statistically significant. ns, not significant; ∗p ≤ 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 4

Trimer supplementation supports siderophore-null UTI89 mutant growth in siderophore-dependent growth medium. Growth of the siderophore-null strain UTI89ΔentBΔybtS was measured by absorbance at 600 nm in siderophore-dependent medium following supplementation with Ent, MGE, or DGE and compared with unsupplemented control (ctrl). Statistics were performed using one-way ANOVA with Dunnett’s multiple-comparison test with p ≤ 0.05 considered as statistically significant. ns, not significant; ∗p ≤ 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Figure 5

Enterobactin (Ent)-associated exometabolites during trimer-dependent growth. Siderophore-null strain UTI89ΔentBΔybtS was cultured in siderophore-dependent medium containing purified Ent, MGE, or DGE. The Ent-associated metabolome in the medium was measured using LC–MS/MS at time points during culture. A–C, Ent (A) is imported and catabolized by UTI89ΔentBΔybtS without producing any dimer (B) or monomer (C) ent catechol compounds. D–F, MGE (D) is imported and catabolized by UTI89ΔentBΔybtS, which produces (DHBS)₂ and G₁-(DHBS)₂ dimers (E) and DHBS monomer (F). G–I, DGE (G) is imported and catabolized by UTI89ΔentBΔybtS, which produces G₁-(DHBS)₂ and G₂-(DHBS)₂ dimers (H) and DHBS monomer (I). Statistics were performed using unpaired t test with p ≤ 0.05 considered as statistically significant. ns, not significant; ∗p ≤ 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Figure 6

Exogenous 2,3-dihydrobenzoic acid (DHB) supports enterobactin (Ent) biosynthesis.A, chemical structure of cyclic Ent with the three DHB-derived groups, containing seven carbon atoms, highlighted in green. B, LC–MS/MS detection of fully ¹³C₃₀-substituted Ent ([M-H]⁻, m/z 698) in UTI89-conditioned ¹³C₃-glycerol culture medium without (−DHB) or with (+DHB), 200 μM unlabeled DHB. C, LC–MS/MS detection of ¹³C₉-substituted Ent ([M-H]⁻, m/z 677) into which three DHB molecules have been incorporated without (−DHB) or with (+DHB), 200 μM unlabeled DHB.

Figure 7

Enterobactin (Ent)-associated dimers support siderophore-dependent growth. A, growth of the siderophore-null strain UTI89ΔentBΔybtS was measured by absorbance at 600 nm in siderophore-dependent medium following supplementation with the Ent-associated dimer exometabolites (DHBS)₂ or G₂-(DHBS)₂ or siderophore-free control (ctrl). B–E, the Ent-associated metabolome in the medium was measured using LC–MS/MS at time points during culture with dimer and monomer results shown for (DHBS)₂-supplemented cultures (B and C) and G₂-(DHBS)₂-supplemented cultures (D and E). Statistics were performed using unpaired t test with p ≤ 0.05 considered as statistically significant. ns, not significant; ∗p ≤ 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. DHBS, DHBS, N-(2,3-dihydroxybenzoyl)serine.