Abstract
A limited understanding of iron uptake mechanisms is available for Streptococcus pyogenes, a haemolytic human pathogen capable of using a variety of haemoproteins in addition to ferric and ferrous iron. This study characterizes a transporter named siu (for streptococcal iron uptake), which consists of an ATP-binding protein (SiuA), a substrate-binding protein (SiuD), and two membrane permease subunits (SiuBG). An siuG mutant was constructed and characterized. The mutant demonstrated growth reduction in comparison to the parent strain when grown in complex medium containing iron in the form of blood, haemoglobin or serum. Only a small reduction in the growth of the siuG mutant was observed in medium containing ferric iron. However, in iron uptake assays the siuG mutant showed a decrease of approximately 30 % in Fe3+ incorporation. Addition of 6 μM haem to the medium inhibited Fe3+ uptake by the wild-type by 76 %, while addition of protoporphyrin IX did not, suggesting that utilization of haem as an iron source is responsible for the inhibition of Fe3+ uptake. Inactivation of siuG moderately reduced the ability of haem to inhibit Fe3+ incorporation by the cells. Inactivation of siaB (encoding a membrane permease of a second iron transporter) had a similar outcome, and inactivation of both transporters had a cumulative effect. These observations implicate both the siu and sia transporters in haem utilization by Strep. pyogenes. Studies in a zebrafish infection model revealed that the siuG mutant was attenuated in both intramuscular and intraperitoneal routes of infection. Together these observations show that the siu system is an iron acquisition pathway in Strep. pyogenes that is important both in vitro and in vivo.
INTRODUCTION
Pathogenic bacteria use a variety of high-affinity iron-scavenging systems to compete for iron while colonizing the human body, as the vast majority of the iron in mammals is tightly bound to host proteins. Most of the intracellular iron is stored in ferritin or found in the form of haem complexes such as haemoglobin and myoglobin. In the body fluids haemoglobin is bound to haptoglobin and haem is carried by haemopexin or serum albumin (Genco & Dixon, 2001; Wandersman & Stojiljkovic, 2000), while ferric iron is sequestered by transferrin and lactoferrin (Wooldridge & Williams, 1993). The redundancy in iron uptake pathways frequently displayed by individual pathogens underscores their importance to the infection process. Accordingly, mutants in one or more of these pathways are often attenuated for virulence (Brown et al., 2001a; Henderson & Payne, 1994; Ratledge, 2004; Stojiljkovic et al., 1995; Torres et al., 2001). Despite its apparent role in bacterial virulence, iron acquisition is only partially understood in Gram-positive pathogens.
The acquisition of iron from host proteins has been studied in a handful of Gram-positive microbes. Some species employ surface receptors for host proteins like transferrin or lactoferrin (Hartford et al., 1993; Modun et al., 1998) or siderophores to obtain ferric iron (Coulanges et al., 1998; Courcol et al., 1997; De Voss et al., 1999; Russell et al., 1984; Sebulsky & Heinrichs, 2001). The use of haem and host haemoproteins has been demonstrated in Corynebacterium diphtheriae (Schmitt, 1997, 1999), Staphylococcus aureus (Mazmanian et al., 2003), and several streptococci (Bates et al., 2003; Brown et al., 2001a; Eichenbaum et al., 1996; Francis et al., 1985; Podbielski et al., 1999). Haem uptake in Gram-positive organisms seems to be mediated by dedicated surface receptors for haem or haemoproteins, while the production of haemophores (secreted haem-binding proteins found in several Gram-negative bacteria) has not been reported (Wandersman & Delepelaire, 2004).
The principal machinery involved in the uptake of free or complex iron in Gram-positive bacteria is ABC transporters, which consist of a substrate-binding lipoprotein, one or two membrane permease subunits, and a hydrophilic ATPase (Brown & Holden, 2002; Gilson et al., 1988; Higgins, 1992; Wandersman & Stojiljkovic, 2000). Haem and siderophore transporters share significant homology and belong to a defined cluster of ABC transporters. In addition, Gram-positive pathogens carry ABC metal transporters, which are part of a separate cluster of transporters and have affinity for multiple metals (Brown & Holden, 2002; Claverys, 2001). In some of these multi-metal transporters, the metal binding receptors function as bacterial adhesins as well (Dintilhac et al., 1997; Elsner et al., 2002; Oligino & Fives-Taylor, 1993; Spellerberg et al., 1999).
Streptococcus pyogenes is a haemolytic pathogen capable of producing a diverse array of skin and mucous membrane infections as well as aggressive deep tissue diseases and streptococcal toxic shock syndrome. Untreated streptococcal infections can lead to the serious complications of rheumatic fever and acute glomerulonephritis (Bisno et al., 2003; Cunningham, 2000). Under laboratory conditions Strep. pyogenes can use haem and a variety of haemoproteins such as haemoglobin-haptoglobin, haemoglobin, myoglobin, haem-albumin and catalase as a source of iron, but it cannot use transferrin or lactoferrin (Eichenbaum et al., 1996; Francis et al., 1985; Podbielski et al., 1999). Strep. pyogenes possesses a multi-metal transporter encoded by mts (Janulczyk et al., 1999) and two transporters from the iron-complex family: sia (streptococcal iron acquisition) (Bates et al., 2003) or hts (Lei et al., 2003) and a transporter which we name here siu (streptococcal iron uptake). The siaABC genes were suggested to function as a haem transporter, and SiaA (or HtsA), the binding protein homologue, was shown to bind haemoglobin and haem (Bates et al., 2003; Lei et al., 2003). On the other hand, the mts transporter is involved in uptake of manganese and ferric iron (Janulczyk et al., 2003), where MtsA binds iron, zinc and manganese in vitro (Janulczyk et al., 1999). The ligand and the function of the siu transporter have not yet been defined. In this study, we investigated the role of the siu transporter in iron acquisition and disease production.
METHODS
Bacterial strains and growth assay conditions.
Escherichia coli DH5α was used for gene cloning. E. coli cells were grown aerobically in Luria–Bertani broth. Strep. pyogenes NZ131 (M49 type; Nordstrand et al., 1998) and the mutant strains were grown statically at 37 °C in Todd–Hewitt (Difco or Oxoid) broth with 0·2 % (w/v) yeast extract (THY), in Todd–Hewitt broth buffered with 10 mM Tris and adjusted to pH 6·9 prior to autoclaving (ZTH), or in ZTH medium containing 12 mM nitrilotriacetic acid (NTA). NTA is a metal chelator with high specificity for iron; its first stability constants (log K1) for Fe3+ and Fe2+ are 15·87 and 8·83 respectively. Since NTA has affinity for zinc, manganese, calcium and magnesium (log K1 of 10·45, 7·44, 6·41 and 5·4 respectively), 0·66 mM of these cations was added to NTA-containing media (Eichenbaum et al., 1996). ZTH was inoculated with a THY overnight culture at a 1/300 dilution. Iron was added to ZTH-NTA in the form of 3 mM ferric chloride (FeCl3), 0·13 % (v/v) sheep's blood (Colorado Serum), 0·13 % (v/v) horse serum (Sigma) or 12 μM human haemoglobin (Sigma). Human haemoglobin was prepared as a 10 mg ml−1 stock solution in phosphate-buffered saline (PBS; pH 7·4).
Todd–Hewitt broth (no yeast extract, TH) was also treated for 3 or 20 h with 5 % (w/v) Chelex-100 (Bio-Rad). The pH of the resin-treated medium was adjusted to 7·65, autoclaved, and 0·5 mM CaCl2 and 0·9 mM MgSO4 were added before inoculation. Inductivity-coupled plasma mass spectrometry (ICP-MS) analysis (at the Laboratory for Environmental Analysis, University of Georgia at Athens) demonstrated that TH contains about 17·5±6·5 μM iron, 0·53±0·2 μM manganese and 15·5±0·2 μM zinc, depending on the batch and manufacturer. Chelex-100 treatment for 3 h resulted in 2·7 μM iron, less than 0·18 μM manganese and about 0·3 μM zinc in the medium. Treatment with Chelex-100 for 20 h did not significantly change its iron content in comparison to 3 h of treatment.
Strep. pyogenes was also grown in a chemically defined medium (CDM) (Podbielski et al., 1999; van de Rijn & Kessler, 1980). CDM was also treated with 3 % (w/v) Chelex-100 for 6 h and filter-sterilized (CxCDM). ICP-MS analysis showed that CxCDM contains 1·6 μM iron and 0·43 μM zinc; the manganese concentration is below the detection level. For cell growth, CxCDM was supplemented with 33 μM MgCl2 and 68 μM CaCl2. All experiments done with Strep. pyogenes cells growing in CDM or CxCDM were inoculated using mid-exponential-phase cells, which were prepared as follows: cells cultured in ZTH medium were harvested at the exponential phase (OD600 0·6), washed twice with PBS, and stored in small frozen aliquots in 16 % (v/v) glycerol. All glassware used for streptococcal growth was soaked for 30 min in a chromic/sulfuric acid solution (Fisher Scientific) and rinsed with double-distilled water (ddH2O). When necessary, the antibiotics spectinomycin and erythromycin were used for E. coli at 100 μg ml−1 and 500 μg ml−1, respectively. For Strep. pyogenes, spectinomycin and erythromycin were used at 100 μg ml−1 and 1 μg ml−1, respectively. Optical density was measured with a Beckman DU640 spectrophotometer (600 nm) or with a Scienceware 800-3 Klett colorimeter (640–700 transmission filter).
Construction of strains ZE4913, ZE4914 and ZE4915.
The primers used in this study are listed in Table 1⇓. Mutants with insertional inactivation of the siuG and siaB genes were constructed in Strep. pyogenes NZ131 (M49 type) using primers designed according to the Strep. pyogenes SF370 genome database (Ferretti et al., 2001) (NCBI and TIGR complete genome databases). All of the constructed chromosomal mutations were verified by PCR analysis. The siuG mutant (ZE4915) was constructed by amplifying a 2·9 kb fragment from NZ131 chromosomal DNA using primers fhuX-S and fhuX-A. The PCR fragment, which included the 3′-end of siuB, the entire siuG gene, and a region downstream of siuG, was digested with AatII and SalI and ligated into pBR322, generating the plasmid pSaAa. The ermAM gene (erythromycin resistance) from pFW15 (Podbielski et al., 1996) was amplified using primers erm-S and erm-A and cloned into the EcoRI site of the siuG gene in pSaAa, generating pSaAaerm. A fragment containing the siuG : : ermAM allele and flanking region was released by AatII/SalI digestion and electroporated into Strep. pyogenes NZ131. Allelic replacement clones were selected on THY agar plates containing erythromycin. A siuG siaB double mutant (ZE4914) was constructed by introducing a disrupted siaB copy in the background of ZE4915. For this purpose, a 2 kb fragment including the 3′-end of siaA, the entire siaB gene and the 5′-end of siaC was amplified from NZ131 chromosomal DNA using primers stoj5 and stoj6. The siaB PCR product was cloned into the XmnI site of pACYC184, producing plasmid pStoj3. The aad9 gene (spectinomycin resistance) was amplified from pUCSpec (Husmann et al., 1997) using spc-S and spc-A primers and cloned into the BclI site of siaB, generating plasmid p5spc1. The fragment containing the siaB : : aad9 allele and flanking chromosomal regions was released by XmnI/StuI digestion and electroporated into the ZE4915 strain. Allelic replacement mutants were selected on THY agar containing erythromycin and spectinomycin. The construction of the siaB mutant (ZE4913) was the same as the construction of ZE4914, except that the siaB : : aad9 allele was introduced into the wild-type NZ131 strain.
Oligonucleotide primer sets
RNA analysis of Strep. pyogenes NZ131 grown in chemically defined medium.
Strep. pyogenes cells were used to inoculate 50 ml CxCDM and CxCDM supplemented with 20 μM FeCl3 or with 33 μM MnSO4 in side-arm (Klett) flasks at 37 °C. Cells were harvested at late-exponential phase (60 Klett units) and RNA was isolated using RiboPure-Bacteria (Ambion), following the manufacturer's recommendations. Reverse transcriptase (RT) reactions were done using Superscript III reverse transcriptase, following the manufacturer's protocol (Invitrogen). Approximately 50 ng of the cDNA was used as the template for each PCR reaction (25 cycles with Taq polymerase, Roche). Gene-specific primer sets listed in Table 1⇑ were used in the RT and PCR reactions at a concentration of 4 μM and 30 μM, respectively.
Ferric iron uptake assays.
Iron uptake assays were essentially performed as previously described, with small modifications (Brown et al., 2001a; Janulczyk et al., 2003). CxCDM was inoculated with Strep. pyogenes cells (1/250) and the cultures were grown to mid-exponential phase (35 Klett units) at 37 °C. 55FeCl3 (0·2 μCi μl−1, 7·4 kBq μl−1; 0·02 μM) was added to 1 ml cultures and incubated at 37 °C. Culture samples (200 μl) were drawn every 30 min and washed twice with 500 μl CxCDM containing 10 mM NTA. The radioactivity associated with the cell pellet and the supernatant was measured as counts per minute (c.p.m.) for 5 min against a 3H standard using a Beckman LS6500 scintillation counter. The culture's OD600 was measured at the same time using a Beckman DU640 spectrophotometer. 55Fe incorporation for each time point was standardized for the cell quantity by dividing the c.p.m. by the culture OD600. Competition assays with iron and manganese were performed as above except that increasing concentrations of non-radioactive FeCl3 or MnSO4 were provided in addition to 55FeCl3. Culture samples were drawn after about 60 min, washed twice, and their radioactivity was measured. For the inhibition of ferric iron uptake, increasing concentrations of haem or protoporphyrin IX were added with the 55FeCl3. Samples were taken after 60 min (OD600 about 1), washed, and the radioactivity measured. 55Fe incorporation was defined as the fraction of c.p.m. of the pellet divided by the sum of the c.p.m. in pellet and the supernatant. Haem was prepared as a 10 mg ml−1 stock solution of haemin chloride (Sigma) in 0·1 M NaOH pH 10. Protoporphyrin IX was prepared as a 10 mg ml−1 stock solution dissolved in 1 : 1 dimethyl formamide/methanol. As a control, a 10 mg ml−1 haem solution was treated with 5 % (w/v) Chelex-100 for 1 h prior to adding it to the cells to remove any unbound iron. When necessary, dilutions were prepared in ddH2O.
Zebrafish care and virulence assays.
Care and feeding of zebrafish (Danio rerio) followed published methods (Neely et al., 2002; Westerfield, 1995). Streptococci were cultured overnight in THY plus 20 % (w/v) peptone (THYP) at 37 °C, diluted 1 : 100 the next day in THYP, and incubated at 37 °C. The cells were harvested at OD600 0·3, washed once with THYP, and diluted to the appropriate concentration in fresh THYP. Injection of zebrafish followed a previously described method (Neely et al., 2002). Briefly, streptococcal cells (10 μl of 105 ml−1) were aseptically injected into groups of four to six anaesthetized male breeder zebrafish (Scientific Hatcheries). Following intraperitoneal (i.p.) or intramuscular (i.m.) injection, the fish were allowed to recover in 225 ml sterilized ddH2O supplemented with aquarium salts (Instant Ocean; Aquarium Systems) at a concentration of 60 mg l−1 in a 29 °C incubator. A control animal group was injected with sterile medium. Infected fish were monitored for 48 h and death recorded in intervals of 12 h. For Strep. pyogenes, the 50 % lethal dose (LD50) for infection of zebrafish was determined by the method of Neely et al. (2002), where zebrafish were challenged over a range of 101–106 c.f.u. of each streptococcal strain.
Zebrafish tissue analysis.
Selected whole zebrafish were fixed following euthanasia at 40 h after infection and 5 μm thick longitudinal sections of the dorsal muscle were prepared for staining as described previously (Neely et al., 2002). Fixed samples were stained with haematoxylin and eosin and examined with an Olympus BX60 microscope equipped with a digital camera and a motorized stage.
Statistics and data analysis.
Statistical significance was determined by using the two-sample Student t-test. The standard error of the mean (sem) was calculated by dividing the standard deviation by the square root of n.
RESULTS AND DISCUSSION
The streptococcal iron uptake (siu) transporter
Sequence analysis of the M1 SF370 Strep. pyogenes genome identified two ABC transporters that belong to the family of iron-complex transporters. One of these streptococcal systems, sia (streptococcal iron acquisition), was previously demonstrated to be involved in haemoglobin binding and acquisition of iron (Bates et al., 2003; Lei et al., 2003). Here we characterize the second iron-complex transporter, which is currently annotated as fhuADB.1G (Spy0383-0386) due to the homology to the corresponding genes of the Bacillus subtilis ferrichrome uptake system. We renamed this transporter siu (streptococcal iron uptake), since we found it to be involved in iron uptake. The siu transporter is conserved in all of the published Strep. pyogenes genomes (five complete and two unfinished genomes). It consists of four genes encoding an ATP-binding protein (siuA), a substrate-binding protein with a lipoprotein signal (siuD), and two subunits of a membrane permease (siuB and siuG). The siuADBG genes constitute an operon that is induced by depletion of iron and other metals (Lei et al., 2003; Smoot et al., 2001). Sequence alignments show that proteins in the Siu system have homology with siderophore uptake systems such as the Staph. aureus SirABC (Heinrichs et al., 1999) and SstABCD (Morrissey et al., 2000) and with transporters involved in haem and haemoprotein utilization such as the staphylococcal HtsABC (Skaar et al., 2004) and IsdDEF (Mazmanian et al., 2003), the pneumococcal PiaABCD and PiuBCDA (Brown et al., 2001a), and the streptococcal SiaABC (Bates et al., 2003).
The siu transporter is involved in iron acquisition
The iron needs of Strep. pyogenes NZ131 were investigated in various media treated with different resins. NZ131 grew in TH medium treated with the chelating resin Chelex-100 or in CxCDM (see Methods), demonstrating that it can proliferate in media containing only 1·6 μM iron and trace amounts of manganese and zinc. Similar observations were made in THY medium that was treated with Chelex-100 (Janulczyk et al., 2003; Ricci et al., 2002). RNA analysis showed that growth in CxCDM allows significant expression of both the siu and the sia transporters (data not shown), which are both negatively regulated by iron (Bates et al., 2003; Lei et al., 2003; Smoot et al., 2001). The addition of 20 μM iron (but not of 20 μM manganese) repressed siu expression, confirming that while the iron concentration found in this medium is sufficient to support growth, it is low enough to produce an iron-stress signal. Growth of NZ131 was significantly impaired in a buffered Todd–Hewitt broth containing 12 mM NTA (ZTH-NTA) that was supplemented with a mix of the bivalent metals calcium, magnesium, manganese and zinc as described by Bates et al. (2003) (black bar, NTA in Fig. 1⇓). Similarly, 20 mM NTA was previously used to restrict the growth of a second M49 strain, CS101, in THY (Podbielski et al., 1999).
Growth analysis of the wild-type and the siuG mutant in iron-depleted media supplemented with various iron sources. Strep. pyogenes wild-type (NZ131, black bars) and siuG (ZE4915, grey bars) cells were grown in iron-depleted medium, ZTH-NTA (ZTH with 12 mM NTA, 0·66 mM MgCl2, MnCl2, CaCl2 and ZnCl2), or in iron-depleted medium supplemented with 3 mM ferric chloride, 0·13 % sheep's blood, 12 μM human haemoglobin or 0·13 % horse serum. The culture's OD600 was determined following overnight incubation. Results are shown as percentage growth with respect to that obtained in complete medium (ZTH). Data represent the mean of at least three experiments. Error bars represent sem; significant P values (<0·05) are indicated by *.
Growth of NZ131 in the NTA-containing medium was significantly enhanced by the addition of different iron sources. The addition of 3 mM ferric chloride, 0·13 % whole blood or 12 μM haemoglobin to ZTH-NTA medium enhanced cell growth up to 86 %, 76 % and 66 %, respectively, of the growth seen in ZTH. On the other hand, the addition of 0·13 % serum resulted in only a small increase of cell growth (32 % of the growth seen in ZTH), indicating that the iron available for the bacterium in serum is low and growth limiting (Fig. 1⇑). These observations suggest that restricted iron availability plays a major role in the restriction of Strep. pyogenes growth in the ZTH-NTA medium. It is possible that this medium limits the availability of other metals as well, since ferric chloride and haemoglobin could not completely restore growth. A mutation in siuG, one of the membrane permease genes, was constructed in Strep. pyogenes NZ131 by the insertional inactivation method using an ermAM cassette (strain ZE4915). The construction of the mutant was verified by PCR analysis. In agreement with the presence of the siuG : : ermAM allele, the PCR product obtained from the siuG mutant was 1·2 kb larger than that obtained from the wild-type strain (data not shown). Growth of the siuG mutant (grey bars in Fig. 1⇑) was compared to that of the parent strain NZ131. The addition of ferric chloride to the iron-depleted medium enhanced the growth of the siuG mutant to a level slightly lower than that of the wild-type (70 % vs 86 % of that observed in ZTH). On the other hand, growth of the siuG mutant was significantly reduced in ZTH-NTA medium containing whole blood or haemoglobin, to only 34 % and 21 %, respectively, of that observed in ZTH. The most significant effect of siuG inactivation was on the ability of the bacteria to grow in ZTH-NTA medium supplemented with serum, where only 3 % of the level of growth observed in ZTH was obtained. Since Strep. pyogenes cannot use transferrin as an iron source, the Siu transporter may contribute to the use of haem bound to serum albumin or haemopexin. While the growth phenotype of the siuG mutant is partial, it establishes the role of the Siu transporter in iron acquisition. Redundancy in iron uptake pathways in Strep. pyogenes is probably a major factor in the partial growth phenotype. However, it is also possible that the siuG mutation allows residual transporter activity (with the remaining membrane permease subunit, siuB), and therefore may hinder the full appreciation of this transporter's role in iron acquisition.
Inactivation of the siu transporter results in decreased Fe3+ utilization
The presence of NTA in the ZTH-NTA medium complicates the study of the cell's use of ferric iron. Therefore, we investigated Fe3+ utilization by the wild-type and the siuG mutant using a 55Fe3+ uptake assay in a low-iron medium that does not contain a chelator. 55FeCl3 was added to cells growing in CxCDM at the early exponential phase and incorporation by the cells was monitored every 30 min. Iron accumulation by both the wild-type and the siuG mutant cells increased over time (Fig. 2⇓), with maximum incorporation by wild-type cells observed after about 30 min incubation. The addition of 2 μM non-radioactive iron (56Fe3+) or manganese (Mn2+) inhibited 55Fe3+ incorporation into NZ131 cells by 30 % and 75 %, respectively. Inhibition reached 62 % and 80 % with 6 μM iron or manganese (data not shown). Inhibition of Fe3+ uptake by manganese suggests that at least some of the Fe3+ uptake in NZ131 is mediated by a multi-metal transporter such as the mts transporter (RT-PCR confirmed the presence of mts transcript in the NZ131 strain; data not shown). Similarly, manganese could compete with Fe3+ uptake in the AP1 (M1 type) strain, in which inactivation of mts reduced accumulation of both 54Mn2+ and 55Fe3+ (Janulczyk et al., 2003).
Incorporation of radioactive ferric iron by the wild-type and the siuG mutant. 55FeCl3 was added to Strep. pyogenes wild-type (NZ131, ▴) and siuG (ZE4915, •) cultures at the early exponential phase grown in CxCDM. Cells were harvested every 30 min, washed, and the radioactivity associated with the cell pellet was measured. Results are expressed as c.p.m. divided by the OD600 of the culture. Data points represent the mean of three experiments done in duplicate; error bars represent sem.
Comparison between the wild-type and the siuG mutant revealed that the rate of Fe3+ uptake was reduced in the mutant, in which maximum incorporation was observed only after 60 min incubation and reached about 70 % of that observed in the wild-type. Therefore, the siu transporter is involved in Fe3+ uptake, although to a lesser extent than the mts transporter, inactivation of which resulted in a 90 % decrease of Fe3+ accumulation (Janulczyk et al., 2003). Similarly to siu, contribution both to Fe3+ uptake and to the cell's ability to grow in iron-restricted medium containing haemoglobin was also demonstrated by the homologous piu and pia transporters in Streptococcus pneumoniae (Brown et al., 2001a).
Haem inhibits Fe3+ accumulation by Strep. pyogenes
We studied the effect of haem present in the growth medium on Fe3+ accumulation by Strep. pyogenes NZ131. Haem at a concentration as low as 0·75 μM inhibits the accumulation of 55Fe3+ by 55 % as compared to the uptake in the absence of haem, and the addition of 6 μM haem results in about 76 % inhibition (Fig. 3a⇓, black bars). Treatment of haem with Chelex-100 to remove free iron possibly present in the solution did not change the percentage inhibition of 55Fe3+ uptake by haem (Fig. 3a⇓, white bar). This is consistent with the proposal that it is the haem, and not free iron, that inhibits 55Fe3+ uptake. On the other hand, the addition of protoporphyrin IX, the core structure of haem, did not significantly interfere with 55Fe3+ incorporation, even at a concentration as high as 6 μM (Fig. 3a⇓, grey bars). Since iron is important for the ability of haem to hinder ferric transport, we suggest that this inhibition is not the outcome of non-specific interference of the haem moiety, but that the use of haem as an iron source leads to repression of the Fe3+ uptake pathways. Haem may also compete with ferric iron for some of the transporters that contribute to Fe3+ uptake.
Inhibition of 55Fe incorporation by haem. 55FeCl3 was added to Strep. pyogenes cultures grown in CxCDM at the early exponential phase. (a) 55Fe accumulation by Strep. pyogenes wild-type (NZ131) in the presence of 6 μM Chelex-treated haem (white bar), or increasing concentrations of haem (black bars), or protoporphyrin IX (grey bars). (b) 55Fe accumulation by the wild-type (NZ131, black bars), the siuG mutant (ZE4915, white bars), the siaB mutant (ZE4913, hatched bars) and the siuG siaB mutant (ZE4914, grey bars) in the presence of increasing concentrations of haem. 55Fe uptake was calculated as incorporation in the presence of haem or protoporphyrin IX as a percentage of the incorporation without the inhibitors. Fe3+ accumulation for each strain was calculated as incorporation in the presence of haem as a percentage of the incorporation in the same strain in the absence of haem. Data represent the mean of at least four experiments; error bars represent sem.
Since we found that haem utilization reduced the accumulation of 55Fe3+ by the wild-type cells, we asked whether inactivation of siuG would interfere with this phenomenon. The ability of haem to hinder 55Fe3+ uptake in the siuG mutant was tested using a range of haem concentrations. The addition of 2 and 4 μM haem resulted in about 54 % and 76 % inhibition of 55Fe3+ uptake in the wild-type strain (Fig. 3b⇑, black bars). However, the ability of haem to inhibit uptake was lessened in the siuG mutant; at 2 and 4 μM haem concentration Fe3+ uptake by the siuG strain was reduced only by 36 % and 60 % (P<0·05, n=4) (Fig. 3b⇑, white bars). Therefore, it seems that the siu is required for effective inhibition of 55Fe3+ transport by haem.
When a higher concentration of haem was used, 55Fe3+ accumulation by the siu mutant was inhibited to a level similar to that observed in the wild-type cells. This suggests that an additional transporter(s) is involved in haem utilization and or is affected by haem. To test this hypothesis, a mutant in siaB (the sia membrane permease component, strain ZE4913) and a siuG siaB double mutant (strain ZE4914) were constructed by the insertional inactivation method using an aad9 cassette. The presence of the siaB : : aad9 allele in both mutant strains was confirmed by PCR; as expected the sia fragment amplified from the siaB and the siuG siaB strains was 1·25 kb larger than that produced from the wild-type strain. To test if the sia transporter contributes to the haem effect seen in the siuG mutant, we repeated the assay with the siaB and the siuG siaB strains. Similar to the siuG mutant, 2 μM haem did not efficiently inhibit 55Fe3+ uptake in the siaB mutant (Fig. 3b⇑, hatched bars, P<0·025, n=6), while 4 μM haem led to a decrease in Fe3+ uptake similar to that seen in the wild-type strain. Inactivation of both siaB and siuG had a cumulative effect at 4 and 6 μM haem. In the presence of 6 μM haem, only 52 % inhibition was observed in the double mutant (Fig. 3b⇑, grey bar), while about 76–80 % inhibition was observed in the wild-type, siuG or siaB strains. While this reduction is not striking, it is statistically significant (P<0·005, n=4) when compared to the wild-type. Based on these observations we suggest that haem utilization is partially impaired if the siu transporter is disrupted and that inactivation of the sia transporter reduces haem usage by the cell even further. Additionally, the ability of haem to reduce 55Fe3+ uptake by 52 % in the double mutant may result from the residual activity of the siu system and the presence of other haem utilization pathways. Redundancy in haem utilization pathways has been demonstrated in several Gram-positive bacteria, such as isdDEF and htsABC in Staph. aureus (Mazmanian et al., 2003; Skaar et al., 2004), piaABCD and piuBCDA in Strep. pneumoniae (Brown et al., 2001a), hmuTUV and an uncharacterized transporter in C. diphtheriae (Drazek et al., 2000; Schmitt & Drazek, 2001).
siuG is required for virulence of Strep. pyogenes in zebrafish
Competitive index studies showed that piaA in S. pneumoniae was important in both a pulmonary and a systemic murine model for disease (Brown et al., 2001a), and mice immunized with recombinant PiuA and PiaA were protected against systemic pneumococcal challenge (Brown et al., 2001b). Likewise, the Strep. pyogenes mtsA and the Staph. aureus hts mutants were attenuated in animal infection models (Janulczyk et al., 2003; Skaar et al., 2004). Using a zebrafish animal model we investigated the role of siuG in disease progression by Strep. pyogenes. The zebrafish immune system has many similarities to the mammalian system (Postlethwait et al., 1998; Trede et al., 2001) and numerous studies have characterized its cardiovascular components (MacRae & Fishman, 2002). Recent studies established that the zebrafish is a suitable model to investigate streptococcal infections. I.p. and i.m. injection of Strep. pyogenes HSC5 (M5 type) produced lethal infections in the fish, along with hypopigmented lesions and tissue necrosis (Miller & Neely, 2004; Neely et al., 2002). We used this model to investigate the role of iron acquisition in disease production and progression by Strep. pyogenes NZ131.
When zebrafish were injected i.m. with a range of 101–106 c.f.u. of the wild-type NZ131, the dose response was similar to that reported for HSC5 (LD50 104 cells ml−1) (Neely et al., 2002). I.m. injection of NZ131 also produced a hypopigmented lesion with extensive muscular necrosis. A control animal group, mock injected with sterile medium, showed no signs of distress. Forty hours after injection, infected zebrafish were fixed and longitudinal sections were prepared. Staining of the tissue revealed streptococcal cells arranged in clusters at the site of infection, as well as the appearance of some host immune cells (data not shown). When the fish were injected i.p. with a range of 101–106 c.f.u. of the wild-type NZ131, the LD50 was higher (>105 cells ml−1) than when they were injected i.m. This observation is different from the observations made with the HCS5 strain, where the LD50 in the i.p. route was lower than that in the i.m. infection (Neely et al., 2002).
We investigated the role of the siu transporter in virulence by comparing the mutant strain to the parent strain when injected separately by both the i.m. and i.p. routes of infection. Groups of four to six zebrafish were challenged with 105 cells ml−1 of Strep. pyogenes wild-type and the siuG mutant and monitored for 2 days. I.m. injection with NZ131 resulted in only 14 % survival of the fish by 48 h (Fig. 4⇓). In the siuG mutant the ability to cause death of the fish was significantly reduced (88 % survival, P<0·0115, n=3). I.p. injection with 105 cells ml−1 of Strep. pyogenes wild-type resulted in about 50 % increase in animal survival as compared to the i.m. injection. Still, the siuG was less virulent as compared to the wild-type (data not shown). These results suggest that acquisition of iron is important for Strep. pyogenes pathogenesis in the zebrafish model and that siuG function has an important role in vivo in the establishment of infection. Iron metabolism and erythroid development in zebrafish is analogous to that of higher vertebrates; zebrafish produce haem and haemoglobin, carry out haemoglobin switching during development (Brownlie et al., 2003), use transferrin receptors (Wingert et al., 2004) and divalent metal transporter 1 (DMT1, Donovan et al., 2002) to transport iron into and within the cell's compartments, and employ ferroprotein 1 (Fpr1) as an intestinal and macrophage iron exporter (Donovan et al., 2000). Therefore, it is likely that the iron acquisition mechanisms used by Strep. pyogenes during infection of zebrafish are relevant for iron acquisition during human infection.
Zebrafish survival curve following Strep. pyogenes infection. Groups of four to six zebrafish were challenged with 105 c.f.u. of Strep. pyogenes wild-type (NZ131, ▴) and the siuG mutant (ZE4915, •) by i.m. injection and monitored for 2 days. Data were pooled from three to five independent experiments and presented as total percentage fish survival as a function of time.
In conclusion, siuADBG is the third ABC-type iron transporter in Strep. pyogenes shown to be involved in iron uptake. The growth assays and the reduced ability of haem to decrease ferric transport by the siuG mutant implicated the siu transporter in iron acquisition from haemoglobin, haem, blood and serum. However, additional studies are required to determine the ligand of the siu transporter and how it contributes to the use of haemoglobin and haem.
Acknowledgments
Griselle Montañez is supported by the predoctoral fellowship 1F31 AI054329-01A1 from the National Institutes of Health (NIH). This work was also supported by the NIH grant RO1 AI057877-01A1. We thank Dr Bernard Beall for the gift of the NZ131 and the ZE4914 strains, Dr Bettina Buttaro for providing the CDM methods, and Dr Erik Skaar for making available the staphylococcal htsABC sequence.