The ssbL gene harbored by the ColV plasmid of an Escherichia coli neonatal meningitis strain is an auxiliary virulence factor boosting the production of siderophores through the shikimate pathway

Abstract

The ability to capture iron is a challenge for most bacteria. The neonatal meningitis Escherichia coli strain S88 possesses several iron uptake systems, notably including siderophores. Transcriptional analysis of the ColV plasmid pS88 has shown strong induction of a previously undescribed gene with low identity to three E. coli chromosomal genes encoding phospho-2-dehydro-3-deoxyheptonate aldolases involved in aromatic amino acid and catecholate/phenolate siderophore biosynthesis through the shikimate pathway. Here, we investigated the role of this gene, ssbLp (ssbL carried on the plasmid), in siderophore biosynthesis and, consequently, in S88 virulence. We constructed an S88 mutant designated S88 ΔssbLp, which exhibited reduced growth under low-iron conditions compared to the wild-type strain. Liquid chromatography-mass spectroscopy analysis of culture supernatants showed that the mutant secreted significantly smaller amounts of enterobactin, salmochelin SX, and yersiniabactin than the wild-type strain. The mutant was also less virulent in a neonatal rat sepsis model, with significantly lower bacteremia and mortality. Supplementation with chorismate, the final product of the shikimate pathway, restored the wild-type phenotype in vitro. In a collection of human extraintestinal E. coli isolates, we found that ssbL was present only in strains harboring the iro locus, encoding salmochelins, and was located either on the chromosome or on plasmids. Acquisition of the iro locus has been accompanied by acquisition of the auxiliary gene ssbL, which boosts the metabolic pathway essential for catecholate/phenolate siderophore biosynthesis and could represent potential therapeutic targets.

FIG 1

Schematic representation of the shikimate pathway and of the biosynthesis of catecholate/phenolate siderophores of Escherichia coli. The first step is an aldol condensation between phosphoenolpyruvate and erythrose-4-phosphate catalyzed by a phospho-2-dehydro-3-deoxyheptonate aldolase (EC 2.5.1.54 according to E. coli nomenclature) to synthesize 3-deoxy-d-arabino-heptulosonate-7-phosphate (DAHP). The last reaction, catalyzed by a chorismate synthase, yields the final product of the pathway, chorismate, which is involved in the biosynthesis of aromatic amino acids and catecholate/phenolate siderophores.

FIG 2

The ssbL_p gene is involved in the ability of E. coli S88 to grow efficiently in iron-depleted medium. Strains were grown in LB medium (A) and in MM9 minimal medium with 20 μM iron (B) or MM9 medium with 0.2 μM iron (C) with and without chorismate supplementation (100 μM). Data presented are averages of results from three independent experiments. *, P < 0.05 compared to other strains or conditions.

FIG 3

LC-MS/MS siderophore profiles of E. coli S88 in iron-poor medium. Enterobactin (A), salmochelin S1 (B), salmochelin SX (C), and aerobactin (D) were extracted after addition of 0.1 M ferric chloride, and cupric yersiniabactin (E) was extracted after addition of 0.5 M copper sulfate. The specific transitions 670 > 224, 627 > 224, 404 > 299, 565 > 205, and 542.9 > 356 m/z were used, respectively, to analyze enterobactin, salmochelin S1, salmochelin X, aerobactin, and cupric yersiniabactin. Data are from the MassLynx software. ES, electrospray positive mode; TIC, total ion current; Sm, smoothed.

FIG 4

Relative production of yersiniabactin, enterobactin, and salmochelin SX measured by LC-MS/MS. Production of yersiniabactin (A), enterobactin (B), and salmochelin SX (C) by the deletion strain S88 ΔssbL_p and the complemented S88 ΔssbLp strain (S88 ΔssbLpC) was compared to that of the wild-type strain S88 in MM9 medium with either 0.2 μM FeSO₄ or 20 μM FeSO₄. Results are expressed in relation to the production of the wild-type strain, which represents 100%. Data are averages of results from six independent experiments and were compared by using Student's t test. Error bars represent the standard deviations. *, P < 0.05; **, P < 0.01.

FIG 5

Lethality of S88, S88 ΔssbL_p, and the complemented S88 ΔssbL_p strain (S88 ΔssbLpC) in a neonatal rat sepsis model. The same inoculum (10³ CFU/ml) was injected intraperitoneally. Levels of bacteremia were compared by using Student's t test, and mortalities were compared with the Gehan-Breslow-Wilcoxon test. *, P < 0.01; **, P < 0.001 (both compared to S88 ΔssbL_p).