Abstract
Monocytes and macrophages are an important host defense in humans infected with Salmonella enterica serovar Typhi. Bacterial ability to survive in these cells is therefore a crucial virulence characteristic of this pathogen. In this study, we demonstrate that growth of a Salmonella enterica serovar Typhi enterochelin synthesis mutant and a tonB mutant in the human monocyte cell line Mono Mac 6 is restricted compared to that of the parental wild-type Ty2 strain. These results suggest that enterochelin- and TonB-mediated iron uptake plays a role in S. enterica serovar Typhi pathogenesis, and also suggest that mutations in iron uptake may attenuate S. enterica serovar Typhi strains for human beings.
1 Introduction
Iron is an essential nutrient for both eukaryotic and prokaryotic cells and plays a central role in electron transport chains involved in intermediary metabolism [1,2]. Because its reactivity under aerobic conditions can generate toxic free radicals, strict control of iron uptake and localization is a necessity. Within mammals for example, levels of free iron are estimated to be less than 10−15 mM because intracellular iron is complexed with heme, Fe–S proteins and ferritin, and extracellular iron is complexed with transferrin in the plasma or with lactoferrin at mucosal surfaces and in exocrine secretions [1,2].
Gram-negative enteric pathogens such as Salmonella enterica synthesize and secrete several classes of low molecular mass Fe+3 chelators (siderophores) such as enterochelin to acquire iron from their mammalian hosts [1–6]. After Fe3+ charging, these siderophores bind to highly specific outer membrane receptors which in turn are transported across the outer membrane into the periplasm using energy provided by the TonB-ExbB-ExbD system [1]. From the periplasm, Fe3+ siderophores are transported across the cell membrane into the cytoplasm by ATP binding cassette-type transport systems [1]. All S. enterica serovar Typhi appear to express enterochelin-mediated iron transport [2]. It is not known whether serovar Typhi expresses an additional but less well-characterized iron uptake system that involves the recently described Sit iron transporter system encoded in the serovar Typhimurium pathogenicity island 1 [7,8].
The ability of S. enterica serovars Typhimurium and Typhi to grow within host cells during systemic infection is an accepted hallmark of bacterial virulence [3,9]. At least some of the non-bactericidal mechanisms by which monocytes and macrophages limit intracellular growth of these and other pathogens include limiting the availability of free iron [1,3,10]. The hostile intracellular environment of the monocyte and macrophage has been shown to upregulate bacterial stress responses modulated by PhoP/PhoQ, OmpR, Fur, and other response regulators [11], again suggesting that the growth of S. enterica serovars under these conditions may be partially restricted by a nutritionally deficient environment. We ourselves have found that serovar Typhi mutants deficient in either enterochelin synthesis or enterochelin-mediated iron transport did not grow well in epithelial cell monolayers or in the presence of unsaturated transferrin or human sera, and were significantly less virulent than wild-type strains in a mouse model of disease [3]. These results in serovar Typhi were in clear contrast to results in serovar Typhimurium, where ent− strains are commonly as virulent for mice as wild-type strains [3,12]. Because we have shown that the enterochelin system is expressed in vivo in patients with typhoid fever [13], we believe that serovar Typhi iron uptake-deficient mutants are likely to be attenuated for human cells. Serovar Typhi aroA mutants are attenuated [14], and, like Escherichia coli, are likely to be deficient in enterochelin synthesis [15]. Because restricted growth in macrophages has been related to attenuation of this organism for human beings [16], we have determined whether the decreased virulence of serovar Typhi iron uptake-deficient mutants in mice [3] is correlated with decreased growth in a non-activated mature human monocytic cell line. We have also completed characterization of the mutation responsible for the defect in enterochelin-mediated transport in serovar Typhi JA055 [3].
2 Materials and methods
2.1 Bacterial strains, cell lines and media
Nalidixic acid-resistant wild-type S. enterica serovar Typhi strain Ty2 and TnPhoA derivative strains JA055, an enterochelin uptake-deficient mutant, and JA138, an ent− mutant, have been previously described [3]. All strains were stored at −20°C in 50% glycerol. Bacteria for macrophage infection were grown overnight in LB medium with 50 μg of nalidixic acid to a concentration of approximately 1.2×109 ml−1, and resuspended to a concentration of 1×109 ml−1 in RPMI 1640 tissue culture medium (Sigma Chemicals, St. Louis, MO, USA). Human monocyte Mono Mac 6 cell line cells [17], a gift from Dr. Michael K. Hoffmann, were cultured at 37°C in 95% air–5% CO2 in RPMI 1640 supplemented with 1 mM sodium pyruvate, 1 mM sodium oxaloacetic acid, non-essential amino acids (Sigma, St. Louis, MO, USA), and 5% fetal calf serum [17]. They were passaged with fresh medium every 3–4 days.
2.2 Identification of TnphoA location in S. enterica serovar Typhi genome
Chromosomal DNA from strain JA055 was digested to completion with BamHI and EcoRI (New England Biolabs). These restriction enzymes cleave within TnphoA to generate fragments containing the kanamycin resistance gene and sequences located downstream (BamHI digestion) and upstream (EcoRI) of the TnphoA insertion. Fragments were cloned into pUC19, recombinant plasmids were electroporated into E. coli DH5α and recombinant clones were selected by kanamycin resistance [18]. The presence of the kanamycin resistance gene in each clone was confirmed by polymerase chain reaction (PCR) using specific primers ((+) 5′-GAAGCGGGAAGGGACTGG-3′; (−) 5′-GTAAAGCACGAGGAAGCGG-3′). Primers for this and other amplifications were chosen using the program Primer 3 Output (Center for Genome Research, Whitehead Institute for Biochemical Research, Cambridge, MA, USA) and synthesized (GenoSys Biotechnology, The Woodlands, TX, USA). Orientation of TnphoA fragments in the recombinant plasmids was determined by restriction analysis using BanII (New England Biolabs) for BamHI recombinant plasmids and SmaI (New England Biolabs) for EcoRI recombinant plasmids.
To determine nucleotide sequences from the junction of the transposon, a pUC19 forward primer ((+) 5′-CGGGCCTCTTCGCTATTACG-3′) and a phoA reverse primer ((−) 5′-GCAGCCCGGTTTTCCAGAAC-3′) were used to amplify the chromosomal inserts in BamHI recombinant plasmids while the same pUC19 forward primer and a IS50R reverse primer ((−) 5′-GAGGTGGTGTCCTCAATGGC-3′) were used to amplify the chromosomal insert in EcoRI recombinant plasmids. Recombinant plasmids containing the EcoRI TnphoA fragment yielded a PCR product of about 3500 bases. The product was purified, and DNA sequenced and analyzed using BLAST as previously described [16], and S. enterica serovar Typhi contigs downloaded from the S. typhi Sequencing Group, Sanger Centre, Cambridge, UK website (ftp://ftp.sanger.ac.uk/pub/pathogens/st).
No PCR product corresponding to the chromosomal insert in BamHI TnphoA recombinant plasmids was obtained, probably because of the large size of the target sequence (>10 000 bases). The BamHI TnphoA fragment was therefore excised from the recombinant plasmid and digested with EcoRV (New England Biolabs) to shorten it; EcoRV does not cut within the TnphoA sequence. BamHI/EcoRV fragments >7000 bases long were subcloned into pBluescript SK(+) (Stratagene). These plasmids were electroporated into E. coli DH5α, and the recombinant plasmids were selected by kanamycin resistance (confirmed by PCR). The predicted single type of recombinant plasmid was obtained, and its identity confirmed by restriction enzyme analysis. Direct pBluescript SK(+) ((+) 5′-GCCAGGGTTTTCCCAGTCAG-3′) and reverse phoA ((−) 5′-GCAGCCCGGTTTTCCAGAAC-3′) primers were used to amplify the chromosomal insert. The PCR product was sequenced and analyzed [18]. The nucleotide sequence of S. enterica serovar Typhi strain JA055 tonB has been deposited in GenBank under accession number AF251127.
2.3 Growth of wild-type serovar Typhi Ty2 and its iron uptake-deficient derivatives in Mono Mac 6 monocytes
Mono Mac 6 cells cultured at 37°C at a concentration of 1×106 cells ml−1 were infected with serovar Typhi strains at a bacteria–monocyte cell ratio of 10:1 for 3 h at 37°C and incubated with 100 μg ml−1 gentamicin for a further 2 h at 37°C to eliminate extracellular bacteria [3]. For morphologic demonstration of the intracellular location of monocyte-associated serovar Typhi, infected monocytes were centrifuged 6 h after treatment with gentamicin onto glass coverslips and stained with Giemsa (Fisher, Springfield, NJ, USA), or with FITC-labeled goat anti-salmonella antibody 3187 (Difco, Detroit, MI, USA) following the manufacturers’ instructions. Coverslips were mounted with Permount and examined under light microscopy or mounted in buffered glycerol and examined with a fluorescence microscope. For quantitation of bacterial growth, aliquots of infected monocyte cultures were taken at 0, 6, 12, and 24 h, monocytes were lysed with 0.1% (v/v) of Triton X-100, the lysates were diluted in phosphate-buffered saline and dilutions were plated in duplicate on LB agar plates. Bacterial colonies were counted after incubation at 37°C for 18 h. Mean and standard error of each point time for eight individual assays were calculated. Intracellular growth was expressed as the ratio between the number of bacteria recovered at time points 6, 12, and 24 h and the number of bacteria recovered at time 0. Data were analyzed statistically by a one-way analysis of variance with a Tukey–Kramer multiple comparison post-test; significance level was set at P <0.05.
3 Results
BLAST analysis of the sequenced PCR-amplified BamHI–EcoRV fragment of strain JA055 revealed a 30-nucleotide sequence identical to a repetitive sequence present in TnphoA (GenBank accession number U25548) fused to a 712-nucleotide sequence that was completely identical to a sequence present in serovar Typhi CT18, 97% identical to serovar Typhimurium tonB (GenBank accession number X56434), and showed lesser degrees of identity to tonB genes from other Gram-negative bacteria. The partial tonB sequence obtained for strain JA055 is identical to a sequence located at nucleotides 1 273 725 to 1 274 436 in the serovar Typhi CT18 genome. The most striking difference between serovar Typhi tonB and serovar Typhimurium tonB was that serovar Typhi JA055 and CT18 tonB sequences had a hexanucleotide reduplication that was not present in serovar Typhimurium tonB, and reciprocally, serovar Typhimurium had a hexanucleotide reduplication that was not present in either Typhi strain tonB. The fusion of TnphoA with tonB began at nucleotide 5 of the tonB open reading frame in strain JA055. The iron uptake defect in strain JA055 is therefore due to transposon-mediated inactivation of tonB.
Strain JA055 is known to contain two TnphoA insertion sequences in its genome [3]. BLAST analysis of other sequenced PCR-amplified EcoRI TnphoA fragments of strain JA055 revealed a sequence 333 nucleotides from the junction of the TnphoA insertion which was identical to a serovar Typhi IS200 insertion sequence (GenBank accession number Y09989), 95–99% identical to other serovar Typhi and Typhimurium IS200 insertion sequences, and identical to a 642-nucleotide sequence in the serovar Typhi CT18 genome. The second insertion in the JA055 genome is therefore located adjacent to an IS200 insertion sequence beginning at nucleotide 659 339.
Both wild-type serovar Typhi strain Ty2 and its iron uptake-deficient tonB (JA055) and ent− (JA138) mutant derivatives invaded human Mono Mac 6 monocytic cells. Intracellular bacteria were readily visible after Giemsa or immunofluorescence staining at 6 h after gentamicin treatment of infected monocytes (Fig. 1). The number of visible intracellular bacteria increased by 12 and 24 h after infection of Mono Mac 6 cells with both wild-type and iron uptake-deficient serovar Typhi strains. Monocytes infected with wild-type serovar Typhi began to lyse 18–24 h after the treatment with gentamicin as a result of proliferation of the intracellular bacteria. Intracellular growth of both serovar Typhi iron uptake-deficient mutants was significantly less than that of wild-type strain Ty2 at 6 h of incubation and beyond (Fig. 2) (P <0.001). By 24 h, the intracellular proliferation of the ent− strain JA138 was significantly less than that of the tonB-inactivated strain JA055 (P <0.01).
Intracellular location of S. enterica serovar Typhi strain JA055 in infected human monocytic Mono Mac 6 cells at 6 h. A: Cells stained with Giemsa stain. Magnification ×630. B: Cells stained with FITC-conjugated anti-Salmonella antibody. Magnification ×630.
Mean intracellular growth (±S.E.M., n=8) of S. enterica serovars Typhi Ty2 (wild-type) (●), JA055 (tonB mutant) (△), and JA138 (ent− mutant) (◻) in human monocytic Mono Mac 6 cells. Mono Mac cells were infected as described in Section 2, lysates of aliquots taken at 0, 6, 12, and 24 h after infection were cultured for 18 h at 37°C, and bacterial colonies counted. Intracellular growth is expressed as the ratio of colony forming units recovered at time t (CFUt) to those recovered at time zero (CFU0).
4 Discussion
Clinical and pathological evidence obtained from patients with typhoid indicates that non-activated monocytes and activated macrophages play an important role in host defenses in both the acute and convalescent stages of S. enterica serovar Typhi infections [3,9]. Improved characterization of the enterochelin uptake-deficient strain JA055 allowed us to determine that it was a tonB rather than a fepA mutant. While both iron uptake mutants and wild-type serovar Typhi strains grew equally well in iron-enriched media or in media with saturated transferrin [3], identification of JA055 as a tonB mutant (this study) and JA138 as an ent− mutant [3], as well as indications that the macrophage phagosome and lysosome have a lowered concentration of iron [10,19], suggests that the human monocytic cell Mono Mac 6 cell line limits the growth of these mutants by iron starvation. Despite the presence of alternative iron uptake systems in S. enterica serovar Typhi [5,7,8,20,21], the inability to take up iron by mechanisms mediated by TonB and enterochelin appears to generate measurable phenotypes of growth restriction in human sera and human cell lines and attenuation of virulence in mice [3].
Our results also suggest that TonB-mediated iron uptake systems and enterochelin in particular may play a particularly important role in the pathogenesis of typhoid fever, and that serovar Typhi mutants such as JA055 and JA138 may be attenuated for human beings. Because JA055 is a tonB mutant, it would be expected to be defective in transport of alternative siderophores used by serovar Typhi such as α-ketoacids and α-hydroxyacids [7,20]. One plausible speculation to account for the different degrees of growth of JA055 and JA138 in macrophages is that enterochelin is the major siderophore of serovar Typhi [22]. Elimination of this siderophore in JA138 leads to greater intracellular growth restriction than does interference with tonB function, because in the latter case, charged enterochelin could be taken up by other as yet uncharacterized TonB-independent mechanisms.
We have previously found that serovar Typhi tonB and ent− mutants are less virulent for mice than the wild-type parental strain [3]. Since serovar Typhi strains attenuated for human beings are known to grow poorly in human macrophages [16], restricted growth of these tonB and ent− mutants in human monocytic cells is consistent with a decrease in virulence. The difference between our findings in serovar Typhi and those of others in serovar Typhimurium [12,21] suggests that iron uptake requirements differ between these two S. enterica serovars. This speculation is consistent with the observation that serovar Typhi strains express lower levels of enterochelin than do serovar Typhimurium strains [22], and that mucin containing iron is needed for serovar Typhi to infect mice [23]. Because serovars Typhi and Typhimurium appear to share similar iron uptake systems, differences in their iron requirements for growth in vivo may reflect differences in the way their iron uptake systems are modulated and expressed under these conditions and/or differences in host-mediated interference with function of these iron uptake systems.
In summary, growth of S. enterica serovar Typhi ent and tonB mutants is restricted in a mature human monocyte cell line. The results of the present and previous studies suggest that despite the complexity of these systems, it is clearly possible to study the role iron uptake systems play in S. typhi virulence using in vitro and in vivo experiments and clinically obtained materials.
Acknowledgements
We would like to thank Mrs. H. Harrison for her help with preparation of the manuscript. This work has been supported by Grant AI43063 (to F.C.C.).
References
