Anaerobic utilization of Fe(III)-xenosiderophores among Bacteroides species and the distinct assimilation of Fe(III)-ferrichrome by Bacteroides fragilis within the genus.

In this study, we show that Bacteroides species utilize Fe(III)-xenosiderophores as the only source of exogenous iron to support growth under iron-limiting conditions in vitro anaerobically. Bacteroides fragilis was the only species able to utilize Fe(III)-ferrichrome while Bacteroides vulgatus ATCC 8482 and Bacteroides thetaiotaomicron VPI 5482 were able to utilize both Fe(III)-enterobactin and Fe(III)-salmochelin S4 as the only source of iron in a dose-dependent manner. We have investigated the way B. fragilis assimilates Fe(III)-ferrichrome as initial model to understand the utilization of xenosiderophores in anaerobes. B. fragilis contains two outer membrane TonB-dependent transporters (TBDTs), FchA1 and FchA2, which are homologues to Escherichia coli ferrichrome transporter FhuA. The disruption of fchA1 gene had only partial growth defect on Fe(III)-ferrichrome while the fchA2 mutant had no growth defect compared to the parent strain. The genetic complementation of fchA1 gene restored growth to parent strain levels indicating that it plays a role in Fe(III)-ferrichrome assimilation though we cannot rule out some functional overlap in transport systems as B. fragilis contains abundant TBDTs whose functions are yet not understood. However, the growth of B. fragilis on Fe(III)-ferrichrome was abolished in a feoAB mutant indicating that Fe(III)-ferrichrome transported into the periplasmic space was reduced in the periplasm releasing ferrous iron prior to transport through the FeoAB transport system. Moreover, the release of iron from the ferrichrome may be linked to the thiol redox system as the trxB deletion mutant was also unable to grow in the presence of Fe(III)-ferrichrome. The genetic complementation of feoAB and trxB mutants completely restored growth on Fe(III)-ferrichrome. Taken together, these findings show that Bacteroides species have developed mechanisms to utilize ferric iron bound to xenosiderophores under anaerobic growth conditions though the regulation and role in the biology of Bacteroides in the anaerobic intestinal environment remain to be understood.

Introduction

The human colon is the most densely populated organ with commensal microbes and Bacteroides species are among the predominant members of that microbiota (Eckburg et al., 2005; Gibson & Roberfroid, 1999; Hooper, Midtvedt, & Gordon, 2002; Savage, 1977). Colonization by Bacteroides spp. is fundamental for the establishment and maintenance of a normal, healthy intestinal microbiota and disruption of this commensal relationship has a great impact on health and disease. In the human colon, Bacteroides spp. can reach numbers in excess of 1011 cells/g of feces and account for about 30–40% of total bacteria where at least 500 different species have been so far reported (Hooper et al., 2002; Smith, Rocha, & Paster, 2006; Xu & Gordon, 2003; Xu et al., 2003). The contribution of this predominant group of bacteria in the large intestine is related to a variety of physiological functions. As an example, Bacteroides spp. are involved directly in complex polysaccharide degradation, bile acid turnover metabolism, proteolytic activity, transformation of toxic and mutagenic compounds, regulation of host fat storage, development of the immune system and protection against pathogens (Eckburg et al., 2005; Jarchum & Pamer, 2011; Neish, 2009; Neu, Douglas‐Escobar, & Lopez, 2007; Reading & Kasper, 2011; Savage, 1977; Smith et al., 2006; Tappenden & Deutsch, 2007).

The diverse bacterial population within the human colon makes this environment a highly competitive ecosystem and in order for Bacteroides spp. to maintain their high cell number, they need to compete efficiently for the available nutrients with other components of the microflora (Fuller & Perdigón, 2003). Among the essential nutrients required by Bacteroides spp. are iron and heme. Bacteroides spp. have an essential requirement for heme and nonheme‐iron and growth can be stimulated in a dose‐dependent manner by heme (Rocha, de Uzeda, & Brock, 1991; Rocha & Smith, 2010; Sperry, Appleman, & Wilkins, 1977; Varel & Bryant, 1974). The Bacteroides are not able to synthesize the tetrapyrrole protoporphyrin IX but can synthesize heme if protoporphyrin IX and a source of inorganic iron is provided in vitro (Rocha & Smith, 2010; Rocha et al., 1991; Sperry et al., 1977). However, there is a paucity of information regarding how Bacteroides species respond to and acquire iron in the anaerobic environment of the human colon. Iron has a remarkable influence on the gut microbiota. The competition for iron fluctuates the balance among commensal bacteria, and iron limitation prevents the colonization of pathogens and mucosa inflammation (Buhnik‐Rosenblau, Moshe‐Belizowski, Danin‐Poleg, & Meyron‐Holtz, 2012; Deriu et al., 2013; Dostal et al., 2012;, Jaeggi et al., 2015; Krebs et al., 2013; Werner et al., 2011; Zimmermann et al., 2010).

Early studies using Enterobacteria as a model have demonstrated that ferrous iron rather than ferric iron was the most important form of iron available to enteric bacteria in the anaerobic environment of the lower intestinal tract (Hantke, 2004; Stojiljkovic, Cobeljic, & Hantke, 1993; Tsolis, Bäumler, Heffron, & Stojiljkovic, 1996). However, recent studies have shown that acquisition of ferric iron via siderophores plays a fundamental role in facultative bacteria colonization of the murine intestinal tract (Pi et al., 2012). In the intestinal tract, ferric iron may be present as insoluble precipitated forms of phytates, carbonates, phosphates, and tannates, and by autooxidation of ferrous iron adjacent to oxygenated mucosal surface (Babbs, 1992; Conrad & Umbreit, 2000). The presence of ferric iron in the colon correlates with recent studies demonstrating that E. coli mono‐ or dual‐associated with Bacteroides thetaiotaomicron in the colonic mucus layer of germ‐free mice induces the expression of genes required for synthesis and uptake of catechol‐type siderophore enterobactin as well as for the uptake of the hydroxamate‐type ferrichrome for the acquisition of ferric iron (Li et al., 2015). These studies indicate that both ferrous and ferric forms of iron are present in the colon but their availability is likely to be limited (Kortman, Raffatellu, Swinkels, & Tjalsma, 2014). Siderophores are low molecular high‐affinity iron chelators synthesized by many microorganisms to forage insoluble ferric iron in aerobic environments or from host tissues iron‐binding proteins when iron availability is limiting (Chu et al., 2010; Ratledge & Dover, 2000).

Aerobic and Facultative Gram‐negative bacteria utilize specific outer membrane TonB‐dependent transporters (TBDTs) to transport iron‐chelates across the outer membrane and into the periplasmic space where periplasmic‐binding proteins and membrane ATP‐binding transporters facilitate their translocation into the cell (Faraldo‐Gómez & Sansom, 2003; Noinaj, Guillier, Barnard, & Buchanan, 2010; Schalk, Mislin, & Brillet, 2012). Transport of substrates through TBDT is energy‐dependent which is derived from the proton motive force and transduced to the outer membrane transporter by the integral inner membrane complex TonB/ExbB/ExbD (Noinaj et al., 2010; Schalk et al., 2012; Schauer, Rodionov, & de Reuse, 2008). Gram‐negative bacteria induce synthesis of TBDTs in response to iron limitation to transport Fe(III)‐siderophores produced by themselves or by other organisms (xenosiderophores) (Armstrong, Brickman, & Suhadolc, 2012; Chu et al., 2010; Galet et al., 2015; Guan, Kanoh, & Kamino, 2001; Joshi, Archana, & Desai, 2006; Krewulak & Vogel, 2008, 2011; Noinaj et al., 2010; Ratledge & Dover, 2000; Strange, Zola, & Cornelissen, 2011; Tanabe et al., 2012). Bacteria also utilize cell‐signaling ECF sigma/antisigma and two‐component regulatory systems to induce the expression of cognate TBDTs in response to the presence of xenosiderophores they are designed to transport (Gasser et al., 2016; Llamas et al., 2006, 2008).

Bacteroides species do not appear to produce known siderophores (Otto, Verweij‐van Vught, van Doorn, & Maclaren, 1988; Rocha et al., 1991) yet they do co‐exist in a habitat densely populated with organisms known to produce siderophores. Therefore, it is reasonable to speculate that Bacteroides could take advantage of xenosiderophores to acquire iron for growth. The Bacteroides robust nutritional versatility is highlighted by the presence of nearly one hundred TBDTs in their genomes which is more than any other bacterium (Cerdeño‐Tárraga et al., 2005; Koebnik, 2005; Patrick et al., 2010; Schauer et al., 2008; Xu et al., 2003). The majority of Bacteroides TBDTs are utilized to import complex polysaccharides and host glycans important for energy generation (Martens, Kelly, Tauzin, & Brumer, 2014; Martens et al., 2011), but for many of these TBDTs receptors the specific substrates and nutritional role remain unknown. Thus in this study, we show that Bacteroides have the capability to grow in the presence of Fe(III)‐xenosiderophores under iron‐limiting conditions anaerobically in vitro. We also show that there is differential assimilation of iron bound to hydroxamate and catechol type siderophores among major Bacteroides species that colonize the human colon. The growth stimulation of B. fragilis by Fe(III)‐ferrichrome and of B. vulgatus and B. thetaiotaomicron by Fe(III)‐enterobactin and Fe(III)‐salmochelin S4 indicates that Bacteroides species have developed significant differences in the way they acquire and compete for iron in the intestinal ecological system.

Materials and methods

Strains and growth conditions

Bacteroides strains and plasmids used in this study are shown in Table 1. Strains were routinely grown anaerobically in brain heart infusion broth supplemented with 5 μg/ml hemin, 1 g/L‐cysteine, and NaHCO3 (BHIS). Rifamycin (20 μg/ml), 100 μg/ml gentamicin, 5 μg/ml tetracycline, and 10 μg/ml erythromycin were added to the media when required. For growth dependence on Fe(III)‐siderophore, a modified semidefined medium (SDM) (Rocha & Smith, 2004) was used as follow: KH2PO4, 1.5 g/L; NH4SO4, 0.5 g/L; NaCl, 0.9 g/L; L‐methionine, 150 mg/L; vitamin B12, 5 μg/L; MgCl2.6H2O, 20 mg/L; CaCl2.2H2O, 10 mg/L; MnCl2.4H2O, 1 mg/L; CoCl2.6H2O, 1 mg/L; resazurin, 1 mg/L; L‐cysteine, 1 g/L; protoporphyrin IX, 5 mg/L; glucose, 5 g/L; tryptone, 1 g/L. Twenty ml of 10% NaHCO3 were added per liter of medium, final pH 7.2. For some experiments, heme was omitted and replaced with protoporphyrin IX (PpIX) as source of tetrapyrrole macrocycle (Rocha et al., 1991). For iron restriction in SDM, the ferrous iron chelator bathophenanthroline disulfonic acid (BPS), which does not enter the cell (Alcaín, Löw, & Crane, 1995; Hassett, Romeo, & Kosman, 1998) was added at 20 μmol/L final concentration. Ammonium ferrous sulfate (Sigma‐Aldrich) at 200 μmol/L was added for ferrous iron‐replete growth conditions. The iron‐free siderophores, ferrichrome (Sigma‐Aldrich), ferrioxamine E (Sigma‐Aldrich), and deferrioxamine (Sigma‐Aldrich) were dissolved in 0.85% sodium chloride and filtered sterilized on 0.20 μm cellulose membrane (Corning Inc., Corning NY). Enterobactin iron‐free (Sigma‐Aldrich), salmochelin S4 iron‐free (Genaxxon bioscience, Germany) and pyoverdine iron‐free (Sigma‐Aldrich) were dissolved in dimethyl sulfoxide (DMSO) and filtered sterilized as above. Iron‐free siderophore stock solutions at 2 mmol/L were mixed with sterile 1 mmol/L ammonium Fe(III) citrate (Sigma‐Aldrich) in distilled water at 1:1 (v/v) overnight to obtain 1 mmol/L siderophore solution with 50% iron‐saturation containing the chelated iron at 0.5 mmol/L in the Fe(III)‐siderophore complexed form. Inoculum cultures were prepared by inoculating 8–10 colonies from fresh cultures grown for 24–48 hr on BHIS plates into 3 ml of SDM broth plus PpIX with addition of 10 μmol/L BPS and incubated anaerobically at 37°C for approximately 24 hr or until reaching OD550 nm of 1.0 allowing exhaustion of cellular endogenous iron. Fresh SDM PpIX media containing 20 μmol/L BPS was inoculated with 1:50 inoculum dilution and supplemented with 0.5 mmol/L Fe(III)‐siderophore solution to final concentrations indicated in the text. For bioassay on BHIS plates, hemin was replaced with 10 μg/ml PpIX and supplemented with 1 mmol/L BPS. A 6 mm sterile disk paper filter was placed on top of inoculated plates and two times 10 μl of 0.5 mmol/L Fe(III)‐siderophore solution was applied on the disk. After 24 hr at 37°C anaerobic incubation, additional two times 10 μl were applied and incubated for 5–6 days.

Table 1

Bacteroides strains and plasmids used in this study

Strains	Relevant genotype	References
B. fragilis 638R	Clinical isolate, Rif	Privitera, Dublanchet, & Sebald, 1979
B. fragilis NCTC 9343	Abdominal infection	NCTC
B. fragilis CLA 267	Clinical isolate Tet Cfx	P. C. Applebauna
B. fragilis IB370	638R trxB::cfxA Rif Cfx	Rocha, Tzianabos, & Smith, 2007
B. fragilis IB383	638R trxB::cfxA pFD892/trxB + Erm	Rocha et al., 2007
B. fragilis BER‐51	638R ΔfeoAB::tetQ, Rif Tet	Veeranagouda et al., 2014
B. fragilis BER‐120	638R fchA2::pFD516 Rif Erm	This study
B. fragilis BER‐125	BER‐51 pER‐191Tet Erm	Veeranagouda et al., 2014
B. fragilis BER‐127	638R fchA1::pYT102 Rif Tet	This study
B. fragilis BER‐128	BER‐127 fchA2::pFD516 Rif Erm Tet	This study
B. fragilis BER‐130	BER‐127 carrying pER‐201 Erm	This study
B. fragilis BER‐131	BER‐128 carrying pER‐201 Erm	This study
B. ovatus ATCC 8483		ATCC
B. thetaiotaomicron VPI 5482		VPI
B. vulgatus ATCC 8482		ATCC
B. vulgatus ATCC 29327		ATCC
B. vulgatus CLA 341		P. C. Applebauna
B. vulgatus 20‐15	Human patient with ulcerative colitis isolate	Onderdonk, Steeves, Cisneros, & Bronson, 1984
B. vulgatus 40G2‐33	Guinea pig with cecal ulceration isolate	Onderdonk et al., 1984
B. vulgatus 10‐9	Health human fecal isolate	Onderdonk et al., 1984
B. vulgatus 16‐4	Health human fecal isolate	Onderdonk, Bronson, & Cisneros, 1987
Plasmids
pYT102	Bacteroides suicide vector, Cm, Tet	Baughn & Malamy, 2002
pFD340	Bacteroides expression shuttle vector, Amp, Erm	Smith, Rogers, & McKee, 1992
pFD516	Bacteroides suicide vector, Sp, Erm	Smith, Rollins, & Parker, 1995
pER‐186	A 0.715 bp BamHI/SstI internal N‐terminus of fchA2 was cloned into the BamHI/SstI sites of pFD516	This study
pER‐178	An approximately 2.4 kb BamHI/EcoRI fragment from pFD340 was deleted and replaced with an approximately 2.4 kb BamHI/EcoRI cfxA gene. Amp Cfx	This study
pER‐194	A 0.604 kb Bamhi/HindIII internal DNA fragment of fchA1 was cloned into the BamHI/HindII sites of pYT102	This study
pER‐201	A 2,522 bp BglII/BamHI promoterless fchA1 DNA fragment was cloned into the BamHI site of pER‐178.	This study

Erm, erythromycin resistance; Cfx, cefoxitine resistance; Rif, rifamycin resistance; Tet, tetracycline resistance; Cm, chloramphenicol resistance; ATCC, American Type Culture Collection; NCTC, National Collection of Type Cultures; VPI, Virginia Polytechnic Institute and State University.

Strain provided by P. C. Applebaun, Department of Pathology, Hershey Medical Center, Pennsylvania 17033.

Construction of B. fragilis fchA1 (BF638R_0018) and fchA2 (BF638R_2503) insertional mutants

An internal 608 nt DNA fragment encompassing nt 64 through 672 of the BF638R_0018 gene locus was PCR amplified using primers Bf‐0018‐BamHI‐Forward (GCCACGGATCCAGAGTCTGTCG) AND Bf‐0018‐HindIII‐Reverse (CTGTAAGCTTTCTACTCCCTGC). The amplified fragment was digested with BamHI/HindIII and cloned into the BamHI/HindIII sites of the E. coli‐Bacteroides shuttle suicide vector pYT102 (Baughn & Malamy, 2002). The new construct, pER‐194, was mobilized from E. coli DH10B into B. fragilis 638R by triparental filter mating protocols previously described (Shoemaker, Getty, Gardner, & Salyers, 1986). Transconjugants were selected on BHIS agar containing 20 μg of rifamycin per ml, 100 μg of gentamicin per ml and 5 μg of tetracycline per ml. PCR amplification analysis was used to confirm single cross‐over insertion of pER‐194 into the new strain BER‐127 (fchA1::pYT102).

An internal 715 nt DNA fragment encompassing nt number 14 through the 729 of the BF638R_2503 ORF was PCR amplified using primers Bf‐2503‐BamHI‐Forward (GAAAAGGATCCTATTAGCTGC) and Bf‐2503‐SstI‐Reverse (CGCGGTGAGCTCCGATACGG). The amplified fragment was digested with BamHI/SstI and cloned into the BamHI/SstI sites of the E. coli‐Bacteroides shuttle suicide vector pFD516 (Smith et al., 1995). The new construct, pER‐186, was mobilized from E. coli DH10B into B. fragilis 638R by triparental filter mating protocols previously described. Transconjugants were selected on BHIS agar containing 20 μg of rifamycin per ml, 100 μg of gentamicin per ml and 10 μg of erythromycin per ml. PCR amplification analysis was used to confirm single cross‐over insertion of pER‐186 into the new strain BER‐120 (fchA2::pFD516).

The construction B. fragilis fchA1 fchA2 double mutant strain (BER‐128) was obtained by mobilizing pER‐194 from E. coli DH10B into BER‐120 strain by triparental mating as described above. Transconjugants were selected on BHIS agar containing 20 μg/ml of rifamycin, 100 μg/ml gentamicin, 10 μg/ml erythromycin and 5 μg/ml tetracycline.

Genetic complementation

For genetic complementation of BER‐127 and BER‐128 strains, a 2,522 nt promoterless DNA fragment of the BF638R_0018 gene locus (fchA1) containing 47 nt upstream the ATG codon was PCR amplified using primers Bf‐0018‐BglII_comp‐Forward (GGTACACAGATCTTTGCGGCTCGC) and Bf‐0018‐BamHI_comp‐Reverse (GCTGATCAGGATCCCTGCCGG) and cloned into the BamHI site of the modified pFD340 (Smith et al., 1992) Bacteroides expression vector pER‐178. The new construct, pER‐201, was conjugated into BER‐127 and BER‐128 by triparental mating to obtain BER‐130 and BER‐131 strains respectively.

Results

Growth stimulation of Bacteroides species by Fe(III)‐siderophores

Fe(III)‐bound siderophores are able to stimulate and support growth of Bacteroides species as the only available form of exogenous iron anaerobically when PpIX was used as the source of tetrapyrrole macrocycle. The B. fragilis 638R, NCTC 9343 and CLA 267 strains were able to grow on solid media around the filter disk loaded with the hydroxamate Fe(III)‐ferrichrome as the only source of iron. None of the other Bacteroides species tested were able to grow on Fe(III)‐ferrichrome (Figure 1 and Table S1). Interestingly, the B. vulgatus and B. thetaiotaomicron strains grew on the catecholates Fe(III)‐enterobactin and Fe(III)‐salmochelin S4 but not on Fe(III)‐ferrichrome (Figure 1 and Table S1). The B. fragilis CLA 267 also grew in the presence of both Fe(III)‐enterobactin and Fe(III)‐salmochelin S4. B. ovatus ATCC 8483 did not grow in the presence of ferrichrome, enterobactin or salmochelin S4. Moreover, none of the Bacteroides strains tested grew in the presence of Fe(III)‐ferrioxamine E, Fe(III)‐pyoverdine, Fe(III)‐deferrioxamine (Table S1) or Fe(III)‐dihydroxybenzoic acid (data not shown). Taken together, these findings indicate that there are significant differences in the assimilation systems of iron‐bound xenosiderophore among intestinal Bacteroides species. No growth was observed around the control plates containing filter disks loaded with ammonium ferric citrate indicating that only iron‐loaded siderophores were able to promote bacterial growth.

Figure 1

Fe(III)‐siderophore‐dependent growth of Bacteroides species on BHIS plates containing 10 μg/ml protoporphyrin IX and 1 mmol/L bathophenanthroline disulfonic acid. Siderophores used are depicted below respective panel. Fe(III)‐siderophore solutions at 0.5 mmol/L were applied onto paper disks placed on top of plates according to the procedures described in the material and methods section. Bacterial colonies grown around the paper disks indicated that growth was stimulated by the addition of respective Fe(III)‐siderophore. Growth control plates contain10 μg/ml protoporphyrin IX plux 200 μmol/L ammonium ferrous sulfate. Abbreviations of the bacterial strains and designations used are labeled beside respective inoculum region in each panel as follow: Bf 638R: B. fragilis 638R, Bf 267: B. fragilis CLA 267, Bo 8483: B. ovatus ATCC 8483, Bt 5482: B. thetaiotaomicron VPI 5482, Bv 8482: B. vulgatus ATCC 8482, Bv 29327: B. vulgatus ATCC 29327, Bv 10‐9: B. vulgatus 10–9, Bv 16‐4: B. vulgatus 16–4, Bv 20‐15: B. vulgatus 20–15, Bv 341: B. vulgatus CLA 341, Bv 40G2‐33: B. vulgatus 40G2‐33

Growth on Fe(III)‐ferrichrome is not totally dependent on the presence of the TonB‐dependent outer membrane proteins FchA1 (BF638R_0018) and FchA2 (BF638R_2503)

As shown in Figure 1, B. fragilis strains were the only Bacteroides species to utilize Fe(III)‐ferrichrome for growth under iron‐limiting conditions on solid agar media. This suggests that a ferrichrome transport system is likely to be present in B. fragilis but absent in other Bacteroides species. This prompted us to characterize the role of putative TBDTs in assimilation of Fe(III)‐ferrichrome.

There are approximately 98 outer membrane TBDTs annotated in the genome of B. fragilis 638R of which 33 are predicted to be TBDTs containing 22 strand antiparallel β‐barrel signature of siderophore transporters and 67 belonging to the SusC family of nutrient transporters ((http://www.ncbi.nlm.nih.gov/nuccore/FQ312004.1; Cerdeño‐Tárraga et al., 2005). A BLAST search revealed two TBDT genes BF638R_0018 and BF638R_2503, henceforth named fchA1 and fchA2 respectively, that are each highly conserved (97%–100% amino acid identity) among all B. fragilis genomes available in the GenBank (83 genomes) but absent in nearly all other Bacteroides species genomes including the B. vulgatus ATCC 8482, B. thetaiotaomicron VPI 5482 and B. ovatus ATCC 8483 strains. These findings correlate with the bioassay results above showing that only B. fragilis was able to grow on Fe(III)‐ferrichrome. FchA1 and FchA2 have amino acid sequence homologies (23% identity, 42% similarity and 24% identity, 41% similarity, respectively) to the well‐studied E. coli Fe(III)‐ferrichrome transporter FhuA (Locher et al., 1998). FchA1 and FchA2 only share 27% identity and 36.5% similarity to each other suggesting that there are significant differences in transport functions and substrate specificities. In fact, an alignment of the amino acid sequence of E. coli FhuA with FchA1 and FchA2 homologs shows that FchA1 but not FchA2 contains a C‐terminal phenylalanine residue important for assembly into the outer membrane (de Cock, Struyvé, Kleerebezem, van der Krift, & Tommassen, 1997; Struyvé, Moons, & Tommassen, 1991). Moreover, the ferrichrome‐binding amino acid residues of E. coli FhuA plug region (R81, G99 and Y116) and barrel region (Y244 and Y315) (Locher et al., 1998) are not conserved in either FchA1 or FchA2 (Figure S1). An unrooted phylogenetic analysis of TBDT homologs in B. fragilis 638R showed that both FchA1 and FchA2 are related to E. coli FhuA in a distinct branch (Figure S2).

These findings prompted us to investigate whether FchA1 and FchA2 would affect growth in the presence of Fe(III)‐ferrichrome. The parent strain, B. fragilis 638R, grew on Fe(III)‐ferrichrome on a dose‐dependent manner while addition of iron in the form of ammonium ferric citrate did not stimulate growth under iron‐limiting conditions (Figures 2a,g). The growth of an fchA1 mutant (BER‐127) and the fchA1 fchA2 double mutant (BER‐128) was partially attenuated when grown on 2 μmol/L and 5 μmol/L Fe(III)‐ferrichrome compared to the parent strain (Figure 2b,e). In contrast, the growth of fchA2 single mutant (BER‐120) was not significantly altered compared to parent strain (Figure 2d). The growth rates of fchA1 or fchA2 mutants were not affected at low concentrations of Fe(III)‐ferrichrome though the genetic complementation of BER‐127 and BER‐128 with the fchA1 gene (BER‐130 and BER‐131 respectively) restored the growth deficiency at 5 μmol/L Fe(III)‐ferrichrome (Figure 2c, f). This suggests that FchA1 is only partly involved in Fe(III)‐ferrichrome utilization. Moreover, the expression of fchA1 and fchA2 mRNAs was not regulated by either inorganic iron‐ or heme‐limiting conditions (Table S2). Taken together, the findings suggest that iron homeostasis is not responsible for control of fchA1 and fchA2 in the uptake and transport of Fe(III)‐ferrichrome in B. fragilis. Despite our limitations in identifying such transporters, we believe that Fe(III)‐ferrichrome utilization in B. fragilis is an active mechanism, and not an artifact effect of growth, because Fe(III)‐ferrichrome has no growth stimulation effect on neither of the related species B. vulgatus, B. thetaiotaomicron nor B. ovatus under the same growth conditions (Figure 1 and 4c–d).

Figure 2

Growth of B. fragilis mutant strains in the presence of Fe(III)‐ferrichrome (a–f) and ammonium Fe(III) citrate (g). Strain designations are depicted in each panel. Bacteria were grown on SDM containing 5 μg/ml protoporphyrin IX and 20 μmol/L bathophenanthroline disulfonic acid. Fe(III)‐ferrichrome (Panels a–f) or ammonium Fe(III) citrate (Panel g) were added at the following final concentrations: No addition (), 0.1 μmol/L (), 0.5 μmol/L (), 2 μmol/L (), and 5 μmol/L (). Ammonium ferrous sulfate at 200 μmol/L () was added for iron‐replete growth controls in all panels. Data presented are an average of two determinations in duplicate (a–f) and one determination in duplicate (g). Vertical bars represent standard deviation. SDM, semidefined medium

The ferrous iron transporter feoAB is required for growth on Fe(III)‐ferrichrome

Interestingly, B. fragilis does not contain homologs of the well‐characterized FhuCDB and FhuF systems of E. coli necessary for binding Fe(III)‐ferrichrome in the periplasmic space, transport across the cytoplasmic membrane, and then reduce it to release ferrous iron in the cytoplasm (Cooper, McArdle, & Raymond, 1978; Fischer, Strehlow, Hartz, & Braun, 1990; Mademidis et al., 1997; Matzanke, Anemüller, Schünemann, Trautwein, & Hantke, 2004). This indicates that assimilation of Fe(III)‐ferrichrome in B. fragilis may differ from the classical mechanism described for facultative Gram‐negative bacteria. We hypothesized that the reduction and release of iron from the Fe(III)‐ferrichrome complex would occur in the periplasm of B. fragilis and the free ferrous iron would be transported into the cytoplasm by the ferrous iron transporter hybrid component system FeoAB (Veeranagouda et al., 2014). To test this, we used the feoAB deletion mutant strain to determine whether it would have growth deficiency in the presence of Fe(III)‐ferrichrome as the only source of exogenous iron. In fact, the feoAB mutant no longer grows on the agar plate in the presence of Fe(III)‐ferrichrome (Figure 3). The genetic complementation of the feoAB with wild‐type gene completely restored the ability of the BER‐51 strain to grow on Fe(III)‐ferrichrome. These findings support our hypothesis that iron released from Fe(III)‐ferrichrome in the periplasmic space is transported into the cytoplasm through the FeoAB system.

Figure 3

Growth deficiency of B. fragilis feo and trxB mutant strains on Fe(III)‐ferrichrome. (a) BHIS plate containing 5 μg/ml protoporphyrin IX plus 200 μmol/L ammonium ferrous sulfate for bacteria growth control. (b) BHIS plates containing 5 μg/ml protoporphyrin IX plus 1 mmol/L bathophenanthroline disulfonic acid as ferrous chelator for exogenous free iron‐limiting conditions. In panel a, sterile saline was added as control of the solvent on the paper disk. In panel b, Fe(III)‐ferrichrome at 0.5 mmol/L solution was added on the paper disk as described in the materials and methods section. Strains designation are labeled in each panel

The mechanism(s) responsible for the reductase activity that causes reduction of ferric iron and its dissociation from ferrichrome in the periplasmic space under anaerobic conditions is not yet known. Nevertheless to investigate whether the redox thiol/disulfate homeostasis in B. fragilis would affect growth on Fe(III)‐ferrichrome, the thioredoxin reductase (TrxB) deletion mutant strain was used (Rocha et al., 2007). Indeed, the trxB mutant was unable to grow around the disk filter containing Fe(III)‐ferrichrome (Figure 3). The genetic complementation of the trxB mutant with wild‐type trxB gene, strain IB383, restored growth on the bioassay plate similar to the parent strain growth (Figure 3). These results clearly indicate that normal physiological redox control is required for this anaerobe to utilize exogenous iron in the form of Fe(III)‐ferrichrome.

Fe(III)‐enterobactin and Fe(III)‐salmochelin S4 support growth of B. vulgatus ATCC 8482

When B. vulgatus ATCC 8482 was cultured in SDM PpIX under iron‐limiting conditions, it grew in the presence of Fe(III)‐enterobactin in a dose‐dependent manner from 0.1 μmol/L to 5 μmol/L. Nearly optimal maximum growth was obtained with 0.5 μmol/L Fe(III)‐enterobactin compared to growth in iron‐replete media (Figure 4a). In contrast, no significant growth occurred when salmochelin S4 was used at 0.1 μmol/L or 0.5 μmol/L. Partial growth occurred at 2 μmol/L while at 5 μmol/L there was a significant growth stimulation reaching maximum growth levels after 24 h compared to iron‐replete conditions (Figure 4b). These findings indicate that Fe(III)‐enterobactin seems to be more efficient in promoting growth of B. vulgatus at lower concentrations than does salmochelin S4 (Figure 4a ,b).

Figure 4

Growth of B. vulgatus ATCC 8482 (a, b, and c) and B. thetaiotaomicron VPI‐5482 (d) on Fe(III)‐siderophores. (a) Fe(III)‐enterobactin. (b) Fe(III)‐salmochelin S4. (c–d) Fe(III)‐ferrichrome. Bacteria were grown on SDM media containing 5 μg/ml protoporphyrin IX and 20 μmol/L bathophenanthroline disulfonic acid. Fe(III)‐siderophores were added at the following concentrations: No addition (), 0.1 μmol/L (), 0.5 μmol/L (), 2 μmol/L (), 5 μmol/L (). Ammonium ferrous sulfate at 200 μmol/L () was added for iron‐replete growth controls. Panels c and d show the growth on Fe(III)‐ferrichrome at 5 μmol/L only for clarity. Data presented are an average of two determinations in duplicate. Vertical bars represent standard deviation. SDM, semidefined medium

The importance of Fe(III)‐enterobactin assimilation in B. vulgatus was further demonstrated for the colitis‐associated B. vulgatus 40G2‐33 and 20‐15 strains in the presence of heme (Figure 5). Interestingly, the B. vulgatus 40G2‐33 and 20–15 strains are unable to grow in the presence of heme at concentrations up to100 μg/ml as the sole source of iron (Figure 5c). However, growth of B. vulgatus 40G2‐33 and 20‐15 can occur in the presence of heme if Fe(III)‐enterobactin (Figure 5b) or inorganic iron are provided exogenously (Figure 5a). In contrast, the control strains B. vulgatus 10–9 and 16–4 isolated from healthy individuals and B. fragilis grew on heme alone as expected (Figure 5c) since in the absence of exogenous iron, iron can be obtained from heme (Rocha et al., 1991; Sperry et al., 1977; Verweij‐Van Vught, Otto, Namavar, Sparrius, & Maclaren, 1988). It is important to mention that growth of Bacteroides species is not stimulated in media lacking heme or protoporhyrin IX (Rocha et al., 1991; Sperry et al., 1977; Verweij‐Van Vught et al., 1988). Taken together, these findings clearly show that intestinal Bacteroides species have developed different strategies to acquire heme‐iron and Fe(III)‐siderophores for growth under iron‐limiting conditions anaerobically.

Figure 5

Growth of B. vulgatus (Bv) and B. fragilis (Bf) strains on BHIS plates supplemented with 100 μg/ml hemin plus (a) 200 μmol/L ammonium ferrous sulfate or b and c) 1 mmol/L bathophenanthroline disulfonic acid. (b) Fe(III)‐enterobactin was added onto the filter disk paper as described in the material and methods section. (c) A solution of 50% DMSO in distilled water was added as onto the disk paper as solvent control. Bacteria strains designation are depicted in each panel

Discussion

In this study, we have demonstrated that Bacteroides species have developed strategies for acquisition of Fe(III)‐bound siderophores produced by other organisms in vitro. The distinct utilization of the hydroxamate Fe(III)‐ferrichrome by B. fragilis and the catecholates enterobactin and salmochelin S4 by B. vulgatus and B. thetaiotaomicron is an advantage for competition for iron and it is likely that it may play a role in growth and composition of the intestinal microflora. The utilization of xenosiderophores by Bacteroides spp. may not only enable them to obtain iron for their own metabolism but also to scavenge intestinal iron as a advantage against competing organisms by limiting environmental iron resources. These findings add further support to reports that commensal microflora play a role in protecting the host against intestinal colonization by pathogenic bacteria by competing and disrupting their ability to forage iron in the lower intestinal tract (Ellermann & Arthur, 2016; Kortman et al., 2014). Very little is known about the mechanisms by which Bacteroides acquire iron in the intestinal tract. We have analyzed the ability of B. fragilis to assimilate iron bound to ferrichrome as an initial model system to understand how anaerobic organisms have developed mechanisms to acquire and utilize ferric iron‐bound siderophores which is a hallmark of iron utilization by aerobic and facultative organisms.

Although progress has been made in recent years in the understanding of the structures and functions of the SusC‐like protein family of TBDTs (Foley, Cockburn, & Koropatkin, 2016; Martens et al., 2011, 2014), very little is known about Bacteroides TBDTs role in the assimilation of iron‐chelate complexes. Bacteroides species contain an extensive number of predicted TBDTs potentially involved in iron acquisition but their substrates and regulatory controls have not been well defined compared to the classical TBDTs in aerobic and facultative bacteria whose cognate substrates and regulation of the transport mechanisms are well understood (Koebnik, 2005; Schalk & Guillon, 2013; Schalk et al., 2012; Schauer et al., 2008). In this study, our efforts to determine the role of FchA1 and FchA2 revealed that only FchA1 plays a role in supporting growth on ferrichrome while FchA2 did not affect growth. In E. coli, deletion of fhuA completely impaired the transport of and growth on Fe(III)‐ferrichrome (Carmel, Hellstern, Henning, & Coulton, 1990). Therefore, we speculate that in view of the abundant number of TBDTs in B. fragilis, it is likely that redundancy in affinity transport of Fe(III)‐ferrichrome is present and highlight the possibility that transport of chelated iron diverges from the eubacterial models.

Another difference between iron utilization in B. fragilis and E. coli is the absence of the periplasmic Fe(III)‐ferrichrome‐binding protein FhuD and of the ATP‐binding cassette transporter FhuBC. In the cytoplasm of E. coli, ferric iron‐bound hydroxamate is released via reduction to ferrous iron with the involvement of the ferric reductase FhuF (Cooper et al., 1978; Matzanke et al., 2004). Mutants defective in FhuE were significantly impaired in their ability to remove iron from coprogen, ferrichrome and ferrioxamine B (Matzanke et al., 2004). In contrast to facultative bacteria, our findings suggest that Fe(III)‐ferrichrome is reduced in the periplasmic space to release free ferrous iron because the inner membrane ferrous iron transporter feoAB mutant has a growth defect when Fe(III)‐ferrichrome is used as the sole source of iron. In this regard, the B. fragilis FeoAB system which is regulated by iron limitation in a classical Fur‐dependent manner (Veeranagouda et al., 2014) may be a major controller of the way B. fragilis regulates the levels of iron that enters cytoplasm to maintain intracellular inorganic iron homeostasis. In support of this, our preliminary studies suggest that this is the case for the removal of iron from heme which also occurs extra‐cytoplasmically and the assimilation of heme‐iron for growth is dependent on the presence of the FeoAB system (unpublished data). This mechanism involving reduction and release of iron from siderophore in the bacterial periplasm has also been shown to occur in Pseudomonas aeruginosa (Greenwald et al., 2007; Marshall, Stintzi, Gilmour, Meyer, & Poole, 2009). Moreover, in the case of ferric iron released from citrate as ferrous iron in the periplasm, it requires the presence of FeoB for transport into the cell (Marshall et al., 2009).

Though B. fragilis has a reducing periplasmic space (Dutton, Boyd, Berkmen, & Beckwith, 2008; Shouldice et al., 2010; Tang, Dallas, & Malamy, 1999), the pathway for ferric iron reductase activities required for Fe(III)‐ferrichrome reduction is unclear. Recent studies have shown that the B. fragilis periplasmic thioredoxin (TrxP) contributes through cycles of reduction and oxidation activities to maintain periplasmic proteins in their reductive state (Shouldice et al., 2010). In B. fragilis, the TrxB/Trx system is the sole mechanism used to maintain the cellular thiol/disulfide balance and the lack of trxB has a major effect on the bacterial growth, oxidative stress response, increased susceptibility to peroxides and thiol oxidants, and survival in intra‐abdominal experimental infections (Reott, Parker, Rocha & Smith, 2009; Rocha et al., 2007). In addition, the TrxB/Trx system seems to be involved in a series of physiological processes in the cytoplasm and periplasm such as the class I aerobic ribonucleotide reductase activity, the protein thiol‐isomerase activities, the periplasm lipoprotein molecular chaperone transport and folding activities (Rocha et al., 2007). Moreover, we cannot rule out at this point of investigation whether the TrxB/Trx redox system may also affect the metal transport activity of the transmembrane hybrid FeoAB fusion system essential for ferrous iron uptake in the Bacteroides (Rocha & Smith, 2010; Veeranagouda et al., 2014).

Although we show here that two major Bacteroides species within the human colon, B. vulgatus and B. thetaiotaomicron, can grow on both Fe(III)‐enterobactin and Fe(III)‐salmochelin S4, the characterization of putative TBDT(s) involved in the catechol transport for these species has not been addressed in this study. Nonetheless, we show in supplemental Figure S3 that in the genome of B. vulgatus ATCC 8482 contain at least thirteen TBDTs homologs to FepA, CirA, and IroN family of enterobactin and salmochelin S4 transporters in Enterobacteria (Müller, Valdebenito, & Hantke, 2009; Schalk & Guillon, 2013). Therefore, it remains to be determined whether these homologs play any role in B. vulgatus utilization of enterobactin and salmochelin S4. Moreover, the absence of significant homologues to periplasmic catechol‐binding protein FepB and ATP‐binding cassette transporter FepCD in the B. vulgatus suggests that the cellular compartment transport and removal of iron from catecholate‐type siderophores may also diverge from aerobic and facultative siderophore transport pathway.

The assimilation of iron bound to enterobactin and salmochelin S4 in Bacteroides may be highly beneficial to the host because it may counteract and neutralize pathogen strategies to evade host defense mechanisms that limit iron in the intestinal tract. One of these strategies is the ability of enteric pathogens to evade the host mucosal secreted antimicrobial glycoprotein lipocalin‐2 (NGAL). Lipocalin‐2 binds Fe(III)‐enterobactin and iron‐free enterobactin disrupting the bacterial iron supply (Flo et al., 2004; Goetz et al., 2002). To circumvent this host defense mechanism, pathogenic enteric bacteria such as S. thyphimurium, Klebsiella pneumonia, uropathogenic E. coli synthesize salmochelin S4, a dual‐glycosylated enterobactin, to by‐pass the lipocalin‐2 inhibitory effect on enterobactin utilization (Hantke, Nicholson, Rabsch, & Winkelmann, 2003; Müller et al., 2009; Smith, 2007; Valdebenito, Müller, & Hantke, 2007). Thus, the ability of B. vulgatus and B. thetaiotaomicron to utilize both enterobactin and salmochelin S4 may disrupt enteric bacteria iron supply by virtue of their sheer numbers as they reach 1011–1012 cfu/g of intestinal content. Again, we think that the differential ability of Bacteroides species to utilize xenosiderophores may not only contribute to competition for iron for their own metabolism and growth, but also protection against mass proliferation of pathogenic organisms in the intestinal tract.

In conclusion, this study shows that Bacteroides species assimilate Fe(III)‐xenosiderophores for growth under anaerobic conditions in vitro. Despite our limited knowledge of iron bound to xenosiderophores assimilation in anaerobes, our findings support previous studies demonstrating that Bacteroides have developed different strategies to deal with the challenges of iron acquisition, genetic regulation and iron‐storage during transitions from anaerobic to aerotolerant metabolism (Betteken, Rocha, & Smith, 2015; Gauss et al., 2012; Rocha & Smith, 2004, 2010, 2013). Moreover, future investigations on the transport and regulatory mechanisms for utilization of catechol siderophores in B. vulgatus associated with colitis will advance our understanding on the role iron acquisition systems play in Bacteroides pathophysiology.

Conflict of Interest

None declared.

Click here for additional data file.