Iron acquisition is an important aspect of the host–pathogen interaction. In the case of Salmonella it is established that catecholate siderophores are important for full virulence. In view of their very high affinity for ferric iron, functional studies of siderophores have been almost exclusively focused on their role in acquisition of iron from the host. In the present study, we investigated whether the siderophores (enterobactin and salmochelin) produced by Salmonella enterica sv. Typhimurium could act as antioxidants and protect from the oxidative stress encountered after macrophage invasion. Our results show that the ability to produce siderophores enhanced the survival of Salmonella in the macrophage mainly at the early stages of infection, coincident with the oxidative burst. Using siderophore biosynthetic and siderophore receptor mutants we demonstrated that salmochelin and enterobactin protect S. Typhimurium against ROS (reactive oxygen species) in vitro and that siderophores must be intracellular to confer full protection. We also investigated whether other chemically distinct siderophores (yersiniabactin and aerobactin) or the monomeric catechol 2,3-dihydroxy-benzoate could provide protection against oxidative stress and found that only catecholate siderophores have this property. Collectively, the results of the present study identify additional functions for siderophores during host–pathogen interactions.

INTRODUCTION

Iron is essential for the growth of almost all living organisms. Although iron is an abundant element, its bioavailability in aerobic environments is limited because of the insolubility of ferric iron. To overcome the difficulty of acquiring sufficient iron for growth, many micro-organisms synthesize a variety of biomolecules (siderophores) that chelate ferric iron and mobilize it for use [1]. Competition for iron is also central to host–pathogen interactions; in the case of Salmonella enterica sv. Typhimurium it has been demonstrated that production of the siderophore salmochelin is essential for full virulence in a murine model of intraperitoneal infection [2]. S. Typhimurium is an intracellular pathogen that replicates in host macrophages. Iron metabolism is important in the macrophage–bacterial interaction; Salmonella infection triggers mobilization of intracellular iron stores through cellular efflux or redistribution to the iron-storage protein ferritin [3]. The observation that increasing iron levels within macrophages leads to increased bacterial load is consistent with the view that withholding iron from the pathogen is critical for defence against infection.

Within the reducing environment of the cell cytoplasm, iron largely exists in the ferrous form. This presents a challenge to cells in an aerobic environment since ferrous iron can react with H2O2 to generate highly toxic hydroxyl radicals via the Fenton reaction [4]. Eukaryotic and prokaryotic cells exert tight control on iron metabolism inside the cell and excess iron is stored to prevent it exerting its pro-oxidant effects on the cell. The fact that ferrous iron can drive the production of hydroxyl radicals has led to the suggestion that iron might be used by the macrophage to potentiate the effects of ROS (reactive oxygen species) that are produced in response to bacterial infection. In this model, iron would be channelled towards the pathogen-containing phagosome [5]. This hypothesis is less favoured at present compared with the ‘iron starvation’ model, in which iron efflux from the phagosome is believed to limit bacterial growth by minimizing acquisition of this essential element [6]. These two models are underpinned by differing views of the mechanism of action of the NRAMP-1 (natural resistance-associated macrophage protein 1) transporter [7]. Against this background we have investigated the role of siderophores during Salmonella–macrophage interaction. In the case of bacterial pathogens the focus on the role of siderophores appears to have been exclusively related to their ability to chelate ferric iron [8]. However, in fungi it has been established that hydroxamate siderophores play an antioxidant role in the cytoplasm as a consequence of their ability to chelate iron [9]. Most studies focused on iron metabolism were conducted on the immortalized cell line RAW264.7, which contains a defective NRAMP system. To provide a more physiologically relevant model, an NRAMP+ macrophage cell line, RAW7.5R, was used in the present study. Our results demonstrate that catecholate siderophores have antioxidant properties and provide new insights into the function of catecholate siderophores in bacterial pathogens.

EXPERIMENTAL

Strains and culture conditions

S. Typhimurium SL1344 [10] and Escherichia coli 83972 [11] were used in the present study. The deletion mutants constructed in these two strains are described in Table 1. The SL1344 entC and iroB mutants were described previously [12]. In the present study, derivatives of these mutants with the kanamycin-resistance gene removed from the chromosome were used. Strains were routinely grown at 37°C in LB broth or MM9 glycerol medium (MM9 medium [13] supplemented with 26.2 mM Mops free acid, 22.1 mM Mops sodium salt, 2 mM MgSO4, 0.1 mM CaCl2, 0.2% glycerol, 0.3% deferrated casamino acids, 0.002% thiamine and 0.2% succinate and adjusted to pH 7). To ensure iron limitation plasticware was used. Where indicated, 50 μM DIP (2,2′-dipyridyl; Sigma) was added to the medium. The S. Typhimurium strains used to infect macrophages were grown in LB medium and harvested as described previously [12].

Table 1
Bacterial strains used in the present study
StrainDescriptionReference
S. Typhimurium and E. coli strains   
 SL1344 wild-type Facultative intracellular pathogen [10
 83972 wild-type Clinical asymptomatic bacteriuria strain [11
Siderophore biosynthetic gene deletion mutants   
 SL1344 S SL1344 ΔiroB The present study 
 SL1344 ES SL1344 ΔentC The present study 
 SL1344 ES* SL1344 ΔentC ΔentE The present study 
 83972 AY 83972 ΔiucABCD ΔybtS; Kanr [18
 83972 ASY 83972 ΔiucABCD ΔiroB ΔybtS; Kanr [18
 83972 EAS 83972 ΔentB ΔiucABCD; Kanr [18
 83972 ESY 83972 ΔentB ΔybtS; Kanr [18
 83972 EASY 83972 ΔentB ΔiucABCD ΔybtS; Kanr [18
Siderophore biosynthetic gene and siderophore receptor gene deletion mutants   
 SL1344 ESR SL1344 ΔfepA ΔiroN The present study 
 SL1344 S-ESR SL1344 ΔiroB ΔfepA ΔiroN The present study 
 SL1344 ES-ESR SL1344 ΔentC ΔfepA ΔiroN The present study 
StrainDescriptionReference
S. Typhimurium and E. coli strains   
 SL1344 wild-type Facultative intracellular pathogen [10
 83972 wild-type Clinical asymptomatic bacteriuria strain [11
Siderophore biosynthetic gene deletion mutants   
 SL1344 S SL1344 ΔiroB The present study 
 SL1344 ES SL1344 ΔentC The present study 
 SL1344 ES* SL1344 ΔentC ΔentE The present study 
 83972 AY 83972 ΔiucABCD ΔybtS; Kanr [18
 83972 ASY 83972 ΔiucABCD ΔiroB ΔybtS; Kanr [18
 83972 EAS 83972 ΔentB ΔiucABCD; Kanr [18
 83972 ESY 83972 ΔentB ΔybtS; Kanr [18
 83972 EASY 83972 ΔentB ΔiucABCD ΔybtS; Kanr [18
Siderophore biosynthetic gene and siderophore receptor gene deletion mutants   
 SL1344 ESR SL1344 ΔfepA ΔiroN The present study 
 SL1344 S-ESR SL1344 ΔiroB ΔfepA ΔiroN The present study 
 SL1344 ES-ESR SL1344 ΔentC ΔfepA ΔiroN The present study 

Infection of mouse macrophages

Murine macrophage-like cells RAW7.5R (NRAMP+ cell line, a gift from Professor Jenefer Blackwell, Telethon Institute for Child Health Research, Perth, WA, Australia) [14] were cultured at 37°C in 5% CO2 in RPMI 1640 medium supplemented with 5% (v/v) FBS and 2 mM L-glutamine. Where indicated, cells were stimulated with 0.5 ng/ml recombinant murine IFNγ (interferon γ; R&D Systems) 16 h before infection. Macrophages were infected with strain SL1344 or SL1344 ES at a MOI (multiplicity of infection) of 10 as described previously [12]. In brief, bacteria were allowed to infect RAW7.5R cells for 1 h prior to removing any extracellular bacteria by incubation with 200 μg/ml gentamicin for 1 h. Subsequently, macrophages were cultured in medium containing 20 μg/ml gentamicin for up to 24 h. Where indicated we supplemented the cell culture medium with 100 μM of the ROS scavenger APDC [(2R,4R)-4-aminopyrrolidine-2,4-dicarboxylic acid, Sigma] 30 min before infection as well as after changing the medium at 1 and 2 h post-infection. To determine the number of intracellular bacteria, macrophages were lysed in PBS supplemented with 0.01% Triton X-100 and the lysate was plated on to LB medium supplemented with 1.5% agar (LB agar). The numbers of intracellular bacteria were calculated by colony counts after an overnight incubation at 37°C. The data presented are the average CFU (colony-forming unit)/ml±S.D. from three independent experiments.

Construction of S. Typhimurium SL1344 iron acquisition deletion mutants and siderophore biosynthetic mutants

To delete the S. Typhimurium SL1344 siderophore receptor genes fepA and iroN, three-step PCR procedures were first employed to generate two amplification products that contained the kanamycin cassette from pKD4 [15] flanked on both sides by approximately 500 bp of DNA sequence homologous with fepA or iroN. The primers used for PCR-mediated gene replacement are listed in Supplementary Table S1 (at http://www.biochemj.org/bj/454/bj4540543add.htm). Siderophore receptor genes were deleted using the λ Red recombinase gene inactivation method as described previously [15]. Plasmid pCP20, expressing the FLP recombinase, was transformed into each deletion mutant to remove the antibiotic resistance cassette prior to subsequent gene deletions. The genotype of all deletion mutants was confirmed by PCR and DNA sequence analysis.

A similar procedure was used to construct the double mutant deficient in entC (isochorismate synthase gene) and entE (enterobactin synthase subunit E gene). The product generated by three-step PCR contained the chloramphenicol cassette from pKD3 [15] flanked on both sides by approximately 500 bp of DNA sequence homologous with entE. The primers used are listed in Supplementary Table S1. The siderophore biosynthetic gene entE was deleted in an entC mutant background using the λ Red recombinase gene inactivation method.

Bacterial growth experiments

For growth assays, strains grown for 16 h in MM9 glycerol medium were used to set up two identical cultures in MM9 glycerol medium at an attenuance at 600 nm (D600) of 0.05. Cultures were then supplemented with either 50 μM DIP or 5 μM FeSO4 and growth was followed by measuring the D600 value. The data presented are representative of three independent experiments.

H2O2 sensitivity on MM9 glycerol agar medium

Bacteria were grown overnight and diluted in MM9 glycerol medium as described above. When the D600 value reached 0.1, aliquots (100 μl) were taken to perform disc diffusion assays on MM9 glycerol medium supplemented with 1.5% agar (MM9 glycerol agar). H2O2 (5 μl of 9.98 M; Sigma) or paraquat (5 μl of 6 M; Sigma) was added to the centre of each plate and allowed to diffuse through the agar. The zones of inhibition were measured after overnight incubation at 37°C. The data presented are representative of three independent experiments. For assays using 2,3-DHB (2,3-dihydroxy-benzoate; Sigma), MM9 glycerol agar supplemented with 10 μg/ml 2,3-DHB was used. To vary the iron concentration in the agar medium, MM9 glycerol agar was either supplemented with 50 μM DIP or left untreated (residual iron). For assays in the presence of exogenous siderophores, overnight cultures of bacteria producing the selected siderophore were first filtered with a 0.2 μm filter. The filtered culture medium, containing secreted siderophores, was then mixed with the bacteria to be seeded on MM9 glycerol agar supplemented with 50 μM DIP.

RESULTS

Survival of wild-type Salmonella and a siderophore biosynthetic mutant in RAW7.5R macrophages

In view of the importance of iron for the intracellular growth of S. Typhimurium in macrophages, we first mutated the entC gene of SL1344 to generate a strain that was unable to synthesize the siderophores enterobactin and salmochelin (referred to as SL1344 ES). Wild-type SL1344 and SL1344 ES were then compared for their ability to survive intracellularly within macrophages. To ensure that iron metabolism in the macrophages followed a pattern that was physiologically relevant, we used the RAW7.5R murine macrophage cell line with a functional NRAMP-1. The RAW7.5R macrophages were untreated or primed with IFNγ 16 h before infection to emphasize the impact of the oxidative burst. Macrophages were then infected with SL1344 or the SL1344 ES mutant and intracellular bacterial loads were determined at 1, 2 and 24 hpi (hours post-infection). In the absence of IFNγ, no significant difference was observed between SL1344 and SL1344 ES intracellular survival at any of the time points (Figure 1A). In IFNγ-primed macrophages, both wild-type SL1344 and SL1344 ES were cleared more efficiently compared with the unprimed macrophages (Figures 1A and 1B), which was expected given the ability of IFNγ to enhance multiple antimicrobial pathways [16]. More importantly, we observed that the number of intracellular SL1344 ES was 20% lower (Student's t test, P<0.05) than wild-type SL1344 at 1 and at 2 hpi, whereas this difference was not apparent at 24 hpi (Figure 1B). Since IFNγ strongly induces NADPH oxidase during the early stages of infection [17], we hypothesized that the increased susceptibility of SL1344 ES is due to the oxidative burst. When the ROS scavenger APDC was co-cultured with IFNγ-primed macrophages before infection, we observed a similar intracellular survival of wild-type SL1344 and SL1344 ES at 1, 2 and 24 hpi (Figure 1C), suggesting that the ability to produce siderophores favours bacterial survival under conditions that promote oxidative burst in macrophages. We also confirmed that iron was limiting in the intramacrophage environment, since iron supplementation of the macrophage culture medium 12 h before infection of primed and unprimed macrophages resulted in enhanced survival of both SL1344 and SL1344 ES at 24 hpi (results not shown). Taken together, these data show that the ability to produce siderophores enhances survival of Salmonella in the macrophage particularly during the early stages of infection. This is the period during which the oxidative burst occurs [17] and it led us to investigate whether the production of siderophores can affect the ability of S. Typhimurium to resist oxidative stress.

Survival of the strains SL1344 and SL1344 ES in RAW7.5R macrophages

Figure 1
Survival of the strains SL1344 and SL1344 ES in RAW7.5R macrophages

Macrophages were left untreated (A), primed with IFNγ 16 h before infection (B) or primed with IFNγ as well as co-cultured with APDC 30 min prior to infection (C). Cells were then infected with strain SL1344 (closed bars) or SL1344 ES (open bars). At 1, 2 and 24 hpi, macrophages were lysed and the CFU were enumerated. Data, displayed as relative survival compared with SL1344 in unprimed macrophages at 1 h, are means±S.D. for three independent experiments. *P<0.05.

Figure 1
Survival of the strains SL1344 and SL1344 ES in RAW7.5R macrophages

Macrophages were left untreated (A), primed with IFNγ 16 h before infection (B) or primed with IFNγ as well as co-cultured with APDC 30 min prior to infection (C). Cells were then infected with strain SL1344 (closed bars) or SL1344 ES (open bars). At 1, 2 and 24 hpi, macrophages were lysed and the CFU were enumerated. Data, displayed as relative survival compared with SL1344 in unprimed macrophages at 1 h, are means±S.D. for three independent experiments. *P<0.05.

Production of siderophores protects S. Typhimurium from ROS

The correlation between the ability to produce siderophores and survival during the period of the oxidative burst suggested that siderophores might influence the ability of S. Typhimurium to resist oxidative stress. To test this hypothesis, the sensitivity of strains SL1344 and SL1344 ES to H2O2 and paraquat, an intracellular superoxide generator, was determined using a disc diffusion assay on MM9 glycerol agar. Figure 2 shows that, for both challenges, the zone of clearance was greater for SL1344 ES compared with SL1344. This is consistent with the view that either salmochelin and/or the non-glycosylated precursor of salmochelin, enterobactin, protect S. Typhimurium against ROS in vitro. To specifically examine the role of salmochelin in this assay an iroB mutant strain (referred to as SL1344 S), that cannot produce salmochelin but still synthesizes enterobactin, was generated. SL1344 S had the same level of resistance to H2O2 and paraquat as SL1344 (Figure 2). This demonstrates that enterobactin is sufficient to protect S. Typhimurium against oxidative stress.

H2O2 and superoxide sensitivities of strains SL1344, SL1344 S and SL1344 ES

Figure 2
H2O2 and superoxide sensitivities of strains SL1344, SL1344 S and SL1344 ES

H2O2 (5 μl of 9.98 M) (A) or paraquat (5 μl of 6 M) (B) was added to the centre of each plate previously seeded with the indicated bacteria. Pictures were taken after an overnight incubation at 37°C. The images presented are representative of results from three independent experiments.

Figure 2
H2O2 and superoxide sensitivities of strains SL1344, SL1344 S and SL1344 ES

H2O2 (5 μl of 9.98 M) (A) or paraquat (5 μl of 6 M) (B) was added to the centre of each plate previously seeded with the indicated bacteria. Pictures were taken after an overnight incubation at 37°C. The images presented are representative of results from three independent experiments.

Since SL1344 ES is unable to synthesize the 2,3-DHB precursor of enterobactin, we tested whether adding this compound to the growth medium of bacteria to be challenged for H2O2 sensitivity could biochemically suppress the phenotype of SL1344 ES. Figure 3 (and Supplementary Figure S1 at http://www.biochemj.org/bj/454/bj4540543add.htm) show that the zone of clearance for SL1344 and SL1344 ES was similar in the presence of 2,3-DHB, indicating that the addition of this biosynthetic intermediate, which restores siderophore production in SL1344 ES, also restores protection against ROS to the levels seen in the wild-type strain. 2,3-DHB is itself a catechol and so to determine whether the protective effect of 2,3-DHB was a consequence of direct action of this molecule or whether it was due to restoration of the production of the catecholate siderophores, we mutated the entE gene in SL1344 ES. The resulting double mutant (referred to as SL1344 ES*) is unable to synthesize 2,3-DHB and also to use externally added 2,3-DHB to synthesize enterobactin. Figure 3 shows that addition of 2,3-DHB did not result in a decreased zone of clearance of SL1344 ES* compared with SL1344 and SL1344 ES. However, SL1344 ES* did show the same resistance to ROS as SL1344 ES in the presence of added siderophore (Supplementary Figure S1). These data show that the presence of the monomeric catechol 2,3-DHB is not sufficient to provide protection against ROS and that 2,3-DHB only provides protection when it is incorporated into the biosynthetic pathway to generate a catecholate siderophore.

Impact of 2,3-DHB supplementation on H2O2 sensitivity of strains SL1344, SL1344 ES and SL1344 ES*

Figure 3
Impact of 2,3-DHB supplementation on H2O2 sensitivity of strains SL1344, SL1344 ES and SL1344 ES*

Bacteria were plated on to either MM9 glycerol agar (open bars) or MM9 glycerol agar supplemented with 10 μg/ml 2,3-DHB (closed bars). Data are displayed as relative zone following exposure to 5 μl of 9.98 M H2O2 of clearance compared with the zones of clearance of SL1344 in the absence of 2,3-DHBs and are means±S.E.M. for three independent experiments.

Figure 3
Impact of 2,3-DHB supplementation on H2O2 sensitivity of strains SL1344, SL1344 ES and SL1344 ES*

Bacteria were plated on to either MM9 glycerol agar (open bars) or MM9 glycerol agar supplemented with 10 μg/ml 2,3-DHB (closed bars). Data are displayed as relative zone following exposure to 5 μl of 9.98 M H2O2 of clearance compared with the zones of clearance of SL1344 in the absence of 2,3-DHBs and are means±S.E.M. for three independent experiments.

The effect of iron on enterobactin-dependent resistance to ROS

The production of enterobactin is associated with conditions of iron limitation and it follows that the effect of this catecholate siderophore on resistance to oxidative killing might be expected to be maximal under the same conditions. On the other hand, intracellular iron can potentiate oxidative killing in the presence of H2O2 and superoxide by promoting Haber–Weiss chemistry. With these scenarios in mind, we compared the H2O2 sensitivity of SL1344 ES and SL1344 S with SL1344 under differing conditions of iron availability. Bacteria were seeded on MM9 glycerol agar supplemented with the iron chelator DIP (50 μM; low iron medium) or left untreated (medium with residual iron) (Figure 4). Differences in iron availability did not impact on the H2O2 sensitivity of SL1344. Strain SL1344 S had a similar sensitivity to H2O2 as SL1344 under both of the conditions tested, consistent with the earlier observation that protection against oxidative stress was mostly dependent on enterobactin. Figure 4 shows that the zone of clearance for SL1344 ES was enhanced by the presence of an iron chelator, but in contrast the zone of clearance was slightly reduced on the medium with residual iron. These results show that enterobactin protects against H2O2 under conditions of iron limitation.

Impact of iron levels on the H2O2 sensitivity of strains SL1344, SL1344 S and SL1344 ES

Figure 4
Impact of iron levels on the H2O2 sensitivity of strains SL1344, SL1344 S and SL1344 ES

To vary the iron concentration, MM9 glycerol agar pates were supplemented with 50 μM DIP (A) or left untreated (B). H2O2 (5 μl of 9.98 M) was added to the centre of each plate previously seeded with the indicated bacteria. Pictures were taken after an overnight incubation at 37°C. The images presented are representative of results from three independent experiments.

Figure 4
Impact of iron levels on the H2O2 sensitivity of strains SL1344, SL1344 S and SL1344 ES

To vary the iron concentration, MM9 glycerol agar pates were supplemented with 50 μM DIP (A) or left untreated (B). H2O2 (5 μl of 9.98 M) was added to the centre of each plate previously seeded with the indicated bacteria. Pictures were taken after an overnight incubation at 37°C. The images presented are representative of results from three independent experiments.

The effect of loss of siderophore receptors on enterobactin-dependent resistance to ROS

To assess whether the protective effect of enterobactin against ROS is linked to iron acquisition, we mutated the siderophore receptor genes for enterobactin (fepA) and salmochelin (iroN) in SL1344, SL1344 S and SL1344 ES. These mutants were designated SL1344 ESR (unable to take up salmochelin/enterobactin), SL1344 S-ESR (unable to synthesize salmochelin or take up salmochelin/enterobactin) and SL1344 ES-ESR (unable to synthesize salmochelin/enterobactin or take up salmochelin/enterobactin). The growth of these mutants was monitored in MM9 glycerol supplemented with 50 μM DIP to ensure iron limitation (Figure 5A). Strains SL1344 ES and SL1344 ES-ESR showed a decrease in exponential growth rates (−52% and −55% respectively compared with SL1344), but reached the same final cell density as SL1344 after 10 h of growth. Strains SL1344 ESR and SL1344 S-ESR possessed a similar initial growth rate as SL1344 ES and SL1344 ES-ESR, but their cell density after 10 h was between 76 and 83% lower than SL1344. This phenotype, which has also been observed with enterobactin and salmochelin receptor mutants in uropathogenic E. coli [18], was attributed to the chelation of iron by the secreted siderophores, which made the metal inaccessible to SL1344 ESR and SL1344 S-ESR. The addition of 5 μM FeSO4 to the culture medium restored the growth of all mutants to the wild-type level, confirming that the growth defects observed were due to iron deficiency (Figure 5B).

Growth of SL1344 wild-type and mutants in MM9 glycerol supplemented with 50 μM DIP (A) or with 5 μM FeSO4 (B)

Figure 5
Growth of SL1344 wild-type and mutants in MM9 glycerol supplemented with 50 μM DIP (A) or with 5 μM FeSO4 (B)

The growth profile of SL1344 ES-ESR, which was the same as SL1344 ES, is not displayed. Growth rates were calculated from the exponential growth phase (0–3.5 h), and cell densities were determined from the 10 h time point. Results are means±S.D. of biological triplicates.

Figure 5
Growth of SL1344 wild-type and mutants in MM9 glycerol supplemented with 50 μM DIP (A) or with 5 μM FeSO4 (B)

The growth profile of SL1344 ES-ESR, which was the same as SL1344 ES, is not displayed. Growth rates were calculated from the exponential growth phase (0–3.5 h), and cell densities were determined from the 10 h time point. Results are means±S.D. of biological triplicates.

The H2O2 sensitivity of the siderophore biosynthesis and receptor mutants was compared using a disc diffusion assay (Figure 6 and Supplementary Figure S2 at http://www.biochemj.org/bj/454/bj4540543add.htm). In this assay, strains SL1344, SL1344 S, SL1344 ESR and SL1344 S-ESR all displayed the same level of sensitivity to H2O2. In contrast, strains SL1344 ES-ESR and SL1344 ES were significantly more sensitive to challenge with H2O2 as indicated by the larger zone of clearance (Supplementary Figure S2). Peripheral growth of these strains outside the zone of clearance indicated that this difference was not due to a growth defect. The effect of H2O2 was apparent over multiple experiments, as represented in Figure 6. Taken together, these data demonstrate that the biosynthesis of enterobactin contributes to resistance to H2O2. Furthermore, the protective effect of enterobactin is not linked to its uptake since the mutant strains that can produce enterobactin, but cannot import it back into the cell (SL1344 ESR and SL1344 S-ESR), still tolerated H2O2 stress better than SL1344 ES.

H2O2 sensitivity of strains with or without the ability to import siderophores

Figure 6
H2O2 sensitivity of strains with or without the ability to import siderophores

The indicated bacteria were plated on to MM9 glycerol agar supplemented with 50 μM DIP. H2O2 (5 μl of 9.98 M) was added to the centre of each plate and the zone of clearance was measured after an overnight incubation at 37°C. The Figure displays data as relative zones of clearance compared with the zone of clearance of SL1344. Results are means±S.E.M. for five independent experiments.

Figure 6
H2O2 sensitivity of strains with or without the ability to import siderophores

The indicated bacteria were plated on to MM9 glycerol agar supplemented with 50 μM DIP. H2O2 (5 μl of 9.98 M) was added to the centre of each plate and the zone of clearance was measured after an overnight incubation at 37°C. The Figure displays data as relative zones of clearance compared with the zone of clearance of SL1344. Results are means±S.E.M. for five independent experiments.

The effect of extracellular siderophores on resistance to ROS

Since strain SL1344 ESR was significantly less sensitive to H2O2 than SL1344 ES-ESR, we then investigated if siderophores presented extracellularly could play a role in H2O2 resistance or if the resistance of SL1344 ESR observed was entirely dependent upon intracellular siderophores not yet secreted. Overnight cultures of SL1344 and SL1344 ES grown in MM9 glycerol medium were filtered using a 0.22 μm filter. The filtered solution containing siderophores secreted by SL1344 was then used to supplement bacteria plated on MM9 glycerol agar containing 50 μM DIP (low iron medium); the filtered solution from SL1344 ES was used in a similar way to a negative control (Supplementary Figure S3 at http://www.biochemj.org/bj/454/bj4540543add.htm). Figure 7 summarizes the H2O2 sensitivity of the strains SL1344, SL1344 ESR, SL1344 ES and SL1344 ES-ESR in the presence or absence of exogenous wild-type siderophores (from the SL1344 filtrate). The addition of salmochelin and enterobactin to SL1344 or SL1344 ESR did not decrease the H2O2 sensitivity of these strains any further. Strain SL1344 ES, which does not produce, but is able to import, siderophores, was rescued by the addition of salmochelin and enterobactin and displayed a wild-type phenotype in the presence of H2O2. Similar conclusions were reached when using an enterobactin solution prepared from an overnight culture of SL1344 S (results not shown). The presence of salmochelin and enterobactin decreased the H2O2 sensitivity of SL1344 ES-ESR; however, this mutant was still more sensitive to H2O2 than wild-type SL1344. These data indicate that extracellular siderophores secreted by Salmonella help to protect the bacteria against oxidative stress, but these siderophores must be intracellular to confer full protection.

Impact of SL1344 siderophore supplementation on H2O2 sensitivity of SL1344, SL1344 ESR, SL1344 ES and SL1344 ES ESR

Figure 7
Impact of SL1344 siderophore supplementation on H2O2 sensitivity of SL1344, SL1344 ESR, SL1344 ES and SL1344 ES ESR

Bacteria plated were either supplemented with the siderophores produced by wild-type SL1344 (closed bars) or the supernatant of a mutant SL1344 ES culture (open bars). Data are displayed as relative zones of clearance compared with the zone of clearance of SL1344 supplemented with wild-type siderophores following exposure to 5 μl of 9.98 M H2O2 and are means±S.E.M. for three independent experiments.

Figure 7
Impact of SL1344 siderophore supplementation on H2O2 sensitivity of SL1344, SL1344 ESR, SL1344 ES and SL1344 ES ESR

Bacteria plated were either supplemented with the siderophores produced by wild-type SL1344 (closed bars) or the supernatant of a mutant SL1344 ES culture (open bars). Data are displayed as relative zones of clearance compared with the zone of clearance of SL1344 supplemented with wild-type siderophores following exposure to 5 μl of 9.98 M H2O2 and are means±S.E.M. for three independent experiments.

Is the protective effect against ROS specific to catecholate siderophores?

In view of our observations with S. Typhimurium, we were interested to determine whether the protective effect of siderophores against oxidative stress was specific to catecholate siderophores or whether other types of siderophore might perform a similar function. Since Salmonella does not produce other types of siderophore, we used the previously constructed and well-defined strains E. coli 83972 EAS (produces yersiniabactin) and 83972 ESY (produces aerobactin) as well as the control strains 83972 ASY (produces enterobactin) and 83972 AY (produces enterobactin and salmochelin) as a source of different siderophores [18]. The siderophore secreted by these strains was collected and used to supplement SL1344 and SL1344 ES to be challenged with H2O2. Figure 8A (and Supplementary Figure S4 at http://www.biochemj.org/bj/454/bj4540543add.htm) show that the H2O2 sensitivity of SL1344 ES supplemented with salmochelin and/or enterobactin produced by E. coli was similar to that of SL1344. However, the addition of yersiniabactin or aerobactin did not rescue SL1344 ES. To confirm this result, we also challenged E. coli 83972 wild-type and siderophore biosynthetic mutants with H2O2 (Figure 8B). The E. coli 83972 mutant unable to produce any siderophore (83972 EASY) was more sensitive to H2O2 than the wild-type strain. A similar phenotype was observed for 83972 ESY (produces aerobactin only) and 83972 EAS (produces yersiniabactin only). In contrast, as observed with Salmonella, E. coli strains that produced enterobactin alone or enterobactin in combination with salmochelin (i.e. 83972 AY and 83972 ASY) were not more affected than the wild-type by the H2O2 challenge. This leads us to conclude that only the catecholate siderophores enterobactin and salmochelin are involved in protection against H2O2.

Protective effect of catecholate and non-catecholate siderophores against H2O2

Figure 8
Protective effect of catecholate and non-catecholate siderophores against H2O2

(A) Impact of catecholate and non-catecholate siderophore supplementation on the H2O2 sensitivity of SL1344 (closed bars) and SL1344 ES (open bars). Bacteria plated were supplemented with the indicated siderophore(s) produced by the E. coli 83972 siderophore biosynthetic mutants mentioned in brackets. Data are displayed as relative zones of clearance compared with the zone of clearance of SL1344 supplemented with wild-type siderophores following exposure to 5 μl of 9.98 M H2O2 and are means±S.E.M. for three independent experiments. (B) H2O2 sensitivity of E. coli 83972 wild-type and siderophore biosynthetic mutants. The indicated strains were seeded on MM9 glycerol agar supplemented with 50 μM DIP. H2O2 (5 μl of 9.98 M) was added to the centre of each plate and pictures were taken after an overnight incubation at 37°C. The images presented are representative of results from three independent experiments.

Figure 8
Protective effect of catecholate and non-catecholate siderophores against H2O2

(A) Impact of catecholate and non-catecholate siderophore supplementation on the H2O2 sensitivity of SL1344 (closed bars) and SL1344 ES (open bars). Bacteria plated were supplemented with the indicated siderophore(s) produced by the E. coli 83972 siderophore biosynthetic mutants mentioned in brackets. Data are displayed as relative zones of clearance compared with the zone of clearance of SL1344 supplemented with wild-type siderophores following exposure to 5 μl of 9.98 M H2O2 and are means±S.E.M. for three independent experiments. (B) H2O2 sensitivity of E. coli 83972 wild-type and siderophore biosynthetic mutants. The indicated strains were seeded on MM9 glycerol agar supplemented with 50 μM DIP. H2O2 (5 μl of 9.98 M) was added to the centre of each plate and pictures were taken after an overnight incubation at 37°C. The images presented are representative of results from three independent experiments.

DISCUSSION

The biochemical mechanism of action of siderophores in iron acquisition is well understood, but there are aspects of their biological roles that are still being delineated. In the case of S. Typhimurium, the observation that production of the glycosylated form of enterobactin (salmochelin) is required for full systemic virulence of Salmonella in a murine model of infection led to an understanding of the interplay between iron acquisition and host innate immune defences [2]. Our initial focus was the influence of siderophores on the survival of S. Typhimurium in macrophages, since it is known that elevated iron levels inside the macrophage are associated with a higher bacterial load [3]. This was also verified with our model (NRAMP+ macrophage RAW7.5R and Salmonella strain SL1344) as addition of iron to the macrophage culture medium led to an increased number of intracellular bacteria at 24 hpi. Our results showed that, when macrophages had been stimulated with IFNγ, there was a significant difference in the recovery of wild-type cells compared with SL1344 ES at 1 and 2 hpi (Figure 1B). IFNγ can influence iron homoeostasis by limiting iron uptake and enhancing iron efflux through ferroportin [19], but it seems unlikely that the decreased survival of the siderophore biosynthetic mutant SL1344 ES was due to an inability to acquire iron since the addition of this metal ion led to a further decrease in bacterial load (results not shown). These results suggested that siderophore production is linked to a function distinct from iron acquisition and, as it occurs at the early stage of infection and is promoted by IFNγ treatment [20], oxidative stress was the likely candidate. Our observation that quenching of ROS production by IFNγ-primed macrophages using APDC led to restoration of SL1344 ES to a bacterial load equivalent to wild-type SL1344 (Figure 1C) is consistent with this view. In the present study there were no growth differences between SL1344 wild-type and mutant strains, except under iron deficiency, and this would be consistent with no gross changes in the transcription of genes associated with defence against oxidative stress.

The involvement of siderophores in tolerance of oxidative stress was confirmed in the in vitro studies that showed that SL1344 ES was highly sensitive to killing by H2O2 and superoxide (Figure 2). Similar observations have been made with Azotobacter vinelandii, where production of the siderophores protochelin and azotochelin provided protection against paraquat [21]. This was attributed to chelation of iron(III) by these siderophores, which prevented oxidative damage catalysed by superoxide and iron(III) [21]. In the case of the fungus Aspergillus fumigatus, siderophore synthesis has also been linked to resistance against oxidative stress as deletion of the gene sidA (L-ornithine N5 mono-oxygenase) produced a siderophore-null mutant with increased H2O2 sensitivity [9]. This increased sensitivity was attributed to an iron deficiency that impaired iron-dependent ROS-detoxifying enzymes such as catalase A. These authors also observed that iron-depleted A. fumigatus was more sensitive to H2O2 than iron-replete cells. In the present study, SL1344 ES grown in the presence of the iron chelator DIP was also more sensitive to H2O2 than under conditions of residual iron (Figure 4). However, since the mutants SL1344 ESR and SL1344 S-ESR (which are unable to import back secreted siderophores) were not more sensitive to H2O2 than SL1344 (Figure 6 and Supplementary Figure S2), we concluded that iron deficiency was not the cause of the increased H2O2 sensitivity of SL1344 ES. It is also worth noting that, although salmochelin is necessary to acquire iron from the host during infection, Salmonella possesses alternative iron uptake systems FeoAB (ferrous iron transport protein AB) [22] and SitABCD (alkaline Mn2+ transporter) [23] that could provide Salmonella with the iron necessary for metalloenzymes in vitro. Our results also showed that extracellular enterobactin provides partial protection, as the H2O2 sensitivity of SL1344 ES-ESR was reduced when extracellular siderophore was present (Figure 7 and Supplementary Figure S3). Since SL1344 ES-ESR was still more sensitive to H2O2 than SL1344 ES when ‘rescued’ with the addition of extracellular enterobactin (and salmochelin), we concluded that the presence of intracellular siderophore was also necessary for maximal protection against H2O2.

S. Typhimurium only produces catechol-type siderophores, but our observation raised the question of whether other chemically distinct siderophores could also provide protection against oxidative stress. We addressed this using a well-characterized set of E. coli mutants that secrete specific siderophores and found that enterobactin, but not yersiniabactin (a four ring structure made up of salicylate, one thiazolidine and two thiazoline rings) or aerobactin (a hydroxamate), could protect against H2O2 stress (Figure 8A and Supplementary Figure S4). This strongly indicates that the antioxidant character of the catechol cannot relate to its iron-binding capability. Enterobactin {N,N′,N″-[(3S,7S,11S)-2, 6, 10-trioxo-1, 5, 9-trioxacyclododecane-3, 7, 11-triyl] tris (2,3dihydroxybenzamide)} is composed of three 2,3-DHB residues connected via an intramolecular scaffold derived from the amino acid serine. We tested the possibility that 2,3-DHB itself could provide protection against oxidative stress. This compound could rescue SL1344 ES, by promoting synthesis of enterobactin, but not SL1344 ES* which is blocked in enterobactin synthesis (Figure 3). These results indicate that enterobactin, but not monomeric catechols, can provide protection. Recently, it has also been observed that enterobactin can provide protection to E. coli against oxidative stress driven by the pseudomonad iron-chelator pyochelin and, again, that other siderophores and the iron chelator citrate could not provide protection [24]. The fate of the catechols is not known, but it seems probable that they form polyphenols which themselves have antioxidant capability. In this context, the similarity to melanins is noted and a wide range of fungal pathogens synthesize melanin to provide protection against oxidative damage. For example, Exophiala dermatitidis, an important fungal pathogen, uses polymerized melanin to absorb the neutrophil oxidative burst [25] and albino mutants of A. fumigatus that are defective in the synthesis of melanin are more susceptible to oxidative killing [26].

Finally, we observed that the H2O2 sensitivity of S. Typhimurium SL1344 ES was decreased under conditions where residual iron was available to the bacteria (Figure 1B). This situation may arise from induction of iron-dependent defences [Bfr (bacterioferritin), SodB (superoxide dismutase) etc.], but our results also suggest that enterobactin-dependent protection against oxidative stress may not be solely linked to iron status. This raises the question of how enterobactin production might be controlled. The first steps of the pathway of enterobactin synthesis is shared with the synthesis of aromatic amino acids, ubiquinone and menaquinone, and p-aminobenzoate (a component of folate) [27]. Chorismate is the common precursor and commitment to production of enterobactin synthesis involves isochorismate hydroxylyase (synthase) encoded by entC [28]. Thus, a potential way in which cells could respond to oxidative stress and increase production of enterobactin would be to channel chorismate towards the siderophore biosynthetic pathway. How this might be achieved in terms of the regulation of enzyme activity and associated signal transduction processes is not yet clear. However, Ma and Payne [29] have reported that an ahpC (alkyl hydroperoxide reductase subunit C) mutant shows a growth defect under conditions of low iron availability and is unable to produce enterobactin. AhpC is an alkyl hydroperoxidase and the two conserved cysteine residues in this protein were observed to be required for the AhpC-dependent production of enterobactin [29]. This leads us to hypothesize that, in addition to its role as a peroxidase protecting the cell against organic peroxides, AhpC may also act as a thiol–disulfide switch sensor that controls flux into the enterobactin biosynthetic pathway in response to oxidative stress at a post-translational level.

Abbreviations

     
  • AhpC

    alkyl hydroperoxide reductase subunit C

  •  
  • APDC

    (2R,4R)-4-aminopyrrolidine-2,4-dicarboxylic acid

  •  
  • CFU

    colony-forming unit

  •  
  • 2,3-DHB

    2,3-dihydroxy-benzoate

  •  
  • DIP

    2,2′-dipyridyl

  •  
  • entC

    isochorismate synthase

  •  
  • entE

    enterobactin synthase subunit E

  •  
  • hpi

    hours post-infection

  •  
  • IFNγ

    interferon γ

  •  
  • NRAMP

    natural resistance-associated macrophage protein

  •  
  • ROS

    reactive oxygen species

AUTHOR CONTRIBUTION

Maud Achard, Rebecca Watts, Matthew Sweet, Mark Schembri and Alastair McEwan conceived the study and designed the experiments; Maud Achard and Kaiwen Chen performed the experiments; Maud Achard, Kaiwen Chen, Matthew Sweet, Mark Schembri and Alastair McEwan analysed the data; Maud Achard, Mark Schembri and Alastair McEwan wrote the paper; and Maud Achard, Kaiwen Chen, Matthew Sweet, Rebecca Watts, Kate Schroder, Mark Schembri and Alastair McEwan revised and performed the editing of the final paper prior to submission.

We thank Professor Jenefer M. Blackwell for the gift of the murine macrophage-like cell line RAW7.5R.

FUNDING

This work was supported by the NHMRC (National Health and Medical Research Council) of Australia [programme grant number 565526 and project grant number APP1005315]. MJS is supported by an ARC (Australian Research Council) Future Fellowship (ID FT100100657) and an honorary NHMRC Senior Research Fellowship (ID APP1003470). MAS is supported by an ARC Future Fellowship (FT100100662).

References

References
1
Sandy
 
M.
Butler
 
A.
 
Microbial iron acquisition: marine and terrestrial siderophores
Chem. Rev.
2009
, vol. 
109
 (pg. 
4580
-
4595
)
2
Crouch
 
M. L. V.
Castor
 
M.
Karlinsey
 
J. E.
Kalhorn
 
T.
Fang
 
F. C.
 
Biosynthesis and IroC-dependent export of the siderophore salmochelin are essential for virulence of Salmonella enterica serovar Typhimurium
Mol. Microbiol.
2008
, vol. 
67
 (pg. 
971
-
983
)
3
Nairz
 
M.
Theurl
 
I.
Ludwiczek
 
S.
Theurl
 
M.
Mair
 
S. M.
Fritsche
 
G.
Weiss
 
G.
 
The co-ordinated regulation of iron homeostasis in murine macrophages limits the availability of iron for intracellular Salmonella typhimurium
Cell. Microbiol.
2007
, vol. 
9
 (pg. 
2126
-
2140
)
4
Imlay
 
J. A.
 
Cellular defenses against superoxide and hydrogen peroxide
Annu. Rev. Biochem.
2008
, vol. 
77
 (pg. 
755
-
776
)
5
Zwilling
 
B. S.
Kuhn
 
D. E.
Wikoff
 
L.
Brown
 
D.
Lafuse
 
W.
 
Role of iron in Nramp1-mediated inhibition of mycobacterial growth
Infect. Immun.
1999
, vol. 
67
 (pg. 
1386
-
1392
)
6
Botella
 
H.
Stadthagen
 
G.
Lugo-Villarino
 
G.
de Chastellier
 
C.
Neyrolles
 
O.
 
Metallobiology of host–pathogen interactions: an intoxicating new insight
Trends Microbiol.
2012
, vol. 
20
 (pg. 
106
-
112
)
7
Nevo
 
Y.
Nelson
 
N.
 
The NRAMP family of metal-ion transporters
Biochim. Biophys. Acta
2006
, vol. 
1763
 (pg. 
609
-
620
)
8
Fischbach
 
M. A.
Lin
 
H. N.
Liu
 
D. R.
Walsh
 
C. T.
 
How pathogenic bacteria evade mammalian sabotage in the battle for iron
Nat. Chem. Biol.
2006
, vol. 
2
 (pg. 
132
-
138
)
9
Schrettl
 
M.
Bignell
 
E.
Kragl
 
C.
Sabiha
 
Y.
Loss
 
O.
Eisendle
 
M.
Wallner
 
A.
Arst
 
H. N.
Haynes
 
K.
Haas
 
H.
 
Distinct roles for intra- and extracellular siderophores during Aspergillus fumigatus infection
PLoS Pathog.
2007
, vol. 
3
 (pg. 
1195
-
1207
)
10
Wray
 
C.
Sojka
 
W. J.
 
Experimental Salmonella Typhimurium infection in calves
Res. Vet. Sci.
1978
, vol. 
25
 (pg. 
139
-
143
)
11
Lindberg
 
U.
Hanson
 
L. A.
Jodal
 
U.
Lidinjanson
 
G.
Lincoln
 
K.
Olling
 
S.
 
Asymptomatic bacteriuria in schoolgirls. II. Differences in Escherichia coli causing asymptomatic and symptomatic bacteriuria
Acta Paediatr. Scand.
1975
, vol. 
64
 (pg. 
432
-
436
)
12
Achard
 
M. E. S.
Tree
 
J. J.
Holden
 
J. A.
Simpfendorfer
 
K. R.
Wijburg
 
O. L. C.
Strugnell
 
R. A.
Schembri
 
M. A.
Sweet
 
M. J.
Jennings
 
M. P.
McEwan
 
A. G.
 
The multi-copper-ion oxidase CueO of Salmonella enterica serovar Typhimurium is required for systemic virulence
Infect. Immun.
2010
, vol. 
78
 (pg. 
2312
-
2319
)
13
Schwyn
 
B.
Neilands
 
J. B.
 
Universal chemical assay for the detection and determination of siderophores
Anal. Biochem.
1987
, vol. 
160
 (pg. 
47
-
56
)
14
Barton
 
C. H.
Whitehead
 
S. H.
Blackwell
 
J. M.
 
Nramp transfection transfers Ity/Lsh/Bcg-related pleiotropic effects on macrophage activation: influence on oxidative burst and nitric oxide pathways
Mol. Med.
1995
, vol. 
1
 (pg. 
267
-
279
)
15
Datsenko
 
K. A.
Wanner
 
B. L.
 
One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products
Proc. Natl. Acad. Sci. U.S.A.
2000
, vol. 
97
 (pg. 
6640
-
6645
)
16
Schroder
 
K.
Hertzog
 
P. J.
Ravasi
 
T.
Hume
 
D. A.
 
Interferon-γ: an overview of signals, mechanisms and functions
J. Leukoc. Biol.
2004
, vol. 
75
 (pg. 
163
-
189
)
17
Vazquez-Torres
 
A.
Jones-Carson
 
J.
Mastroeni
 
P.
Ischiropoulos
 
H.
Fang
 
F. C.
 
Antimicrobial actions of the NADPH phagocyte oxidase and inducible nitric oxide synthase in experimental salmonellosis. I. Effects on microbial killing by activated peritoneal macrophages in vitro
J. Exp. Med.
2000
, vol. 
192
 (pg. 
227
-
236
)
18
Watts
 
R. E.
Totsika
 
M.
Challinor
 
V. L.
Mabbett
 
A. N.
Ulett
 
G. C.
De Voss
 
J. J.
Schembri
 
M. A.
 
Contribution of siderophore systems to growth and urinary tract colonization of asymptomatic bacteriuria Escherichia coli
Infect. Immun.
2012
, vol. 
80
 (pg. 
333
-
344
)
19
Nairz
 
M.
Fritsche
 
G.
Brunner
 
P.
Talasz
 
H.
Hantke
 
K.
Weiss
 
G.
 
Interferon-γ limits the availability of iron for intramacrophage Salmonella typhimurium
Eur. J. Immunol.
2008
, vol. 
38
 (pg. 
1923
-
1936
)
20
VanderVen
 
B. C.
Yates
 
R. M.
Russell
 
D. G.
 
Intraphagosomal measurement of the magnitude and duration of the oxidative burst
Traffic
2009
, vol. 
10
 (pg. 
372
-
378
)
21
Cornish
 
A. S.
Page
 
W. J.
 
The catecholate siderophores of Azotobacter vinelandii: their affinity for iron and role in oxygen stress management
Microbiology
1998
, vol. 
144
 (pg. 
1747
-
1754
)
22
Jeon
 
J.
Kim
 
H.
Yun
 
J.
Ryu
 
S.
Groisman
 
E. A.
Shin
 
D.
 
RstA-promoted expression of the ferrous iron transporter FeoB under iron-replete conditions enhances Fur activity in Salmonella enterica
J. Bacteriol.
2008
, vol. 
190
 (pg. 
7326
-
7334
)
23
Zhou
 
D. G.
Hardt
 
W. D.
Galan
 
J. E.
 
Salmonella typhimurium encodes a putative iron transport system within the centisome 63 pathogenicity island
Infect. Immun.
1999
, vol. 
67
 (pg. 
1974
-
1981
)
24
Adler
 
C.
Corbalan
 
N. S.
Seyedsayamdost
 
M. R.
Pomares
 
M. F.
de Cristobal
 
R. E.
Clardy
 
J.
Kolter
 
R.
Vincent
 
P. A.
 
Catecholate siderophores protect bacteria from pyochelin toxicity
PLoS One
2012
, vol. 
7
 pg. 
e46754
 
25
Schnitzler
 
N.
Peltroche-Llacsahuanga
 
H.
Bestier
 
N.
Zundorf
 
J.
Lutticken
 
R.
Haase
 
G.
 
Effect of melanin and carotenoids of Exophiala (Wangiella) dermatitidis on phagocytosis, oxidative burst, and killing by human neutrophils
Infect. Immun.
1999
, vol. 
67
 (pg. 
94
-
101
)
26
Jahn
 
B.
Koch
 
A.
Schmidt
 
A.
Wanner
 
G.
Gehringer
 
H.
Bhakdi
 
S.
Brakhage
 
A. A.
 
Isolation and characterization of a pigmentless-conidium mutant of Aspergillus fumigatus with altered conidial surface and reduced virulence
Infect Immun.
1997
, vol. 
65
 (pg. 
5110
-
5117
)
27
Dosselaere
 
F.
Vanderleyden
 
J.
 
A metabolic node in action: chorismate-utilizing enzymes in microorganisms
Crit. Rev. Microbiol.
2001
, vol. 
27
 (pg. 
75
-
131
)
28
Walsh
 
C. T.
Liu
 
J.
Rusnak
 
F.
Sakaitani
 
M.
 
Molecular studies on enzymes in chorismate metabolism and the enterobactin biosynthetic pathway
Chem. Rev.
1990
, vol. 
90
 (pg. 
1105
-
1129
)
29
Ma
 
L.
Payne
 
S. M.
 
AhpC is required for optimal production of enterobactin by Escherichia coli
J. Bacteriol.
2012
, vol. 
194
 (pg. 
6748
-
6757
)

Supplementary data