Iron transport has been linked to the virulence of Brucella strains in both natural and experimental hosts. The genes designated BAB2_0837–0840 in the Brucella abortus 2308 genome sequence are predicted to encode a CupII-type ferrous iron transporter homologous to the FtrABCD transporter recently described in Bordetella. To study the role of the Brucella FtrABCD in iron transport, an isogenic ftrA mutant was constructed from B. abortus 2308. Compared with the parental strain, the B. abortusftrA mutant displays a decreased capacity to use non-haem iron sources in vitro, a growth defect in a low iron medium that is enhanced at pH 6, and studies employing radiolabelled FeCl3 confirmed that FtrABCD transports ferrous iron. Transcription of the ftrA gene is induced in B. abortus 2308 in response to iron deprivation and exposure to acid pH, and similar to other Brucella iron acquisition genes that have been examined the iron-responsiveness of ftrA is dependent upon the iron response regulator Irr. The B. abortusftrA mutant exhibits significant attenuation in both cultured murine macrophages and experimentally infected mice, supporting the proposition that ferrous iron is a critical iron source for these bacteria in the mammalian host.
Brucella strains cause abortion and infertility in their natural animal hosts. Humans can also acquire a chronic, debilitating febrile illness known as brucellosis, or ‘undulant fever’, as the result of contact with infected animals or their products (Pappas et al., 2006). Like most bacteria, Brucella strains require iron as an essential micronutrient (Waring et al., 1953). Acquiring enough iron to meet their physiological needs is a particular challenge for these bacteria because they live almost exclusively in association with a mammalian host. During infection, the brucellae live mainly inside macrophages, and their capacity to survive and replicate within these cells allows them to produce chronic disease, which is the hallmark of human brucellosis (Roop et al., 2009). The activation of host macrophages by IFN-γ generally limits iron availability to bacteria such as the brucellae that live within phagosomal compartments in these phagocytes. This occurs through the reduced production of transferrin receptors on the surface of activated macrophages (Byrd and Horwitz, 1989) and the stimulation of the natural resistance-associated macrophage protein 1 (Nramp1) to export divalent cations including iron out of the phagosomal compartment (Cellier et al., 2007). IFN-γ also enhances the expression of the iron exporter ferroportin 1 resulting in increased iron efflux out of macrophages (Nairz et al., 2008). The extent to which these individual host iron deprivation strategies influences the availability of this essential micronutrient to Brucella strains during their intracellular replication in host macrophages, however, remains unknown.
The brucellae have evolved multiple strategies for overcoming iron limitation in their mammalian hosts, including siderophore-dependent ferric iron-specific and haem acquisition systems (Bellaire et al., 2003; Paulley et al., 2007). However, no ferrous iron (Fe2+)-specific transport system has been previously described in Brucella strains, and no genes have been discovered in the currently available Brucella genome sequences that are predicted to encode the Feo- or Yfe/Sit-type ferrous iron transporters that are found in many other bacterial pathogens (Roop et al., 2011). The identification and characterization of a ferrous iron transporter in Brucella would represent an important finding because during the early stages of their intracellular residence in host macrophages the brucellae reside in acidic compartments known as endolysosomal Brucella-containing vesicles (eBCVs) (Starr et al., 2012), where Fe2+ could potentially be a biologically relevant iron source (Roop et al., 2011).
Recently, several bacterial ferrous iron transporters has been identified that rely upon homologues of the eukaryotic iron permease Ftr1p (Stearman et al., 1996) for their activity. A subset of these transporters exemplified by the EfeUOB transporter of Escherichia coli (Große et al., 2006; Cao et al., 2007) and the FtrABCD transporter of Bordetella (Brickman and Armstrong, 2012) possess proteins with predicted cupredoxin (Cup) domains that are thought to convert Fe2+ to Fe3+ prior to transport by their Ftrp1 homologues (Rajasekaran et al., 2010) in a process analogous to the coupled function of the ferroxidase Fet3p and Ftr1p in Saccharomyces cerevesiae (De Silva et al., 1995; Stearman et al., 1996). The proposed cupredoxin domains of the putative ferroxidases in these transporters (e.g. EfeO and FtrB) have different structural features (Fig. S2), which has led to the designation of EfeUOB and homologous systems as Cup-I type ferrous iron transporters and FtrABCD and homologous systems as Cup-II type ferrous iron transporters (Rajasekaran et al., 2010). The genes designated BAB2_0837–0840 in the Brucella abortus 2308 genome sequence are predicted to encode the components of a Cup-II-type ferrous iron transporter. The purpose of the studies described in this report was to evaluate the role of this transporter in iron transport and virulence in Brucella.
Discovery of an Ftr1-based ferrous iron transporter in B. abortus 2308
A survey of the B. abortus 2308 genome sequence revealed four genes designated BAB2_0837–0840 predicted to encode the components of an iron transporter belonging to the Cup-II-type family of ferrous iron transporters (Fig. 1), and homologous to the recently described Bordetella iron transporter FtrABCD (Brickman and Armstrong, 2012). Correspondingly, ftrA (BAB2_0840) is predicted to encode a homologue of the periplasmic iron-binding protein known as P19 in some bacteria (Chan et al., 2010), ftrB (BAB2_0839) a Cup-II-type ferroxidase, ftrC (BAB2_0838) a homologue of the iron permease Ftr1 found in eukaryotic microbes such as S. cerevesiae (Stearman et al., 1996) and Candida albicans (Ramanan and Wang, 2000), and ftrD (BAB2_0837) a polyferredoxin (Fig. 1). Consistent with their proposed functions, the Brucella FtrA protein possesses the conserved metal-binding domains that have been proposed to be important for the iron-binding properties of P19 proteins in other bacteria (Fig. S1) (Chan et al., 2010), FtrB contains the conserved H106 and D103 residues and RKEKV and EXE domains that have been proposed to be critical for the ferroxidase activity of the bacterial cupredoxins involved in Ftr1-mediated iron transport (Fig. S2) (Rajasekaran et al., 2010), and FtrC contains the two conserved RExxE motifs required for the iron permease activity of Ftr1 (Fig. S3) (Fang and Wang, 2002; Severance et al., 2004). FtrD also exhibits extensive amino acid homology with other bacterial ‘ferrodoxins’ such as the Bordetella FtrD protein which have been linked to the activity of Ftr1-based ferrous iron transporters, including the presence of multiple cysteine rich motifs which likely co-ordinate the 4Fe-4S clusters (Fig. S4) required for the proposed electron transport activity of these proteins (Rajasekaran et al., 2010). SignalP and PSORT analysis of the amino acid sequences of the Brucella proteins suggest that FtrA is exported to the periplasm, FtrB is anchored to the outer surface of the cytoplasmic membrane and that FtrC and D are integrated into the cytoplasmic membrane, all of which are cellular locations consistent with the proposed roles of these proteins in the import of Fe2+.
A B. abortusftrA mutant exhibits iron acquisition defects in vitro
To test if the FtrABCD transporter is required for the growth of B. abortus 2308 under iron-limited conditions, we compared the growth of the parental 2308 strain and the ftrA mutant in the low-iron minimal medium originally described by López-Goñi et al. (1992). Although both of these strains exhibited comparable growth profiles in this medium during an initial round of culture (data not shown), the ftrA mutant began to lose viability during stationary phase during a second round of growth in this medium (Fig. 2A). This phenotype was not observed for the parental 2308 strain nor a derivative of the ftrA mutant in which the ftrA locus had been reconstructed (Fig. 2A). Moreover, the ftrA mutant did not exhibit this loss of stationary phase viability after an initial round of growth in low iron minimal medium followed by re-inoculation into this medium supplemented with FeCl3 (Fig. 2B). Whether this phenotype is the result of an enhanced need for iron during stationary phase or simply a reflection of a general depletion of cellular iron stores is unknown. The ftrA mutant also displayed reduced zones of growth compared with the parental strain around two ‘non-haem’ iron sources [FeCl3 and Fe(SO4)2(NH4)2] in a solid medium-based iron source utilization assay (Fig. 3), but all three of these strains showed an equivalent capacity to use haem as an iron source. These experimental findings support the proposition that FtrABCD is an iron transporter in B. abortus 2308.
FtrABCD is a ferrous iron transporter
Although the B. abortus ftrA mutant exhibited a defect in its capacity to use both ferric chloride (FeCl3) and ferrous ammonium sulphate [Fe(SO4)2(NH4)2] in the solid medium-based disk diffusion assay, this type of assay is not typically used for differentiating between the ability of a bacterial strain to transport Fe2+ or Fe3+. Indeed, considering the concentrations of the iron sources added to the disks, the defect in the capacity of the ftrA mutant to use FeCl3 as an iron source could be a reflection of the fact that this strain, unlike the parental strain, cannot utilize the low concentration of Fe2+ that exists in equilibrium with Fe3+ under these experimental conditions. Consequently, in an attempt to gain better insight into specificity of the FtrABCD transporter, we compared the growth of the parental 2308 strain and ftrA mutant in brucella broth buffered at pH 6 or 7 in presence of the Fe3+-specific chelator desferroxamine. The ftrA mutant exhibited a significant growth defect compared with the parental 2308 strain under iron-restricted conditions at pH 6, but not at pH 7 (Fig. 4). Given that acidic pH favours the stability of Fe2+ (Crichton, 2009), these results provided indirect evidence that FtrABCD is an Fe2+ transporter.
To confirm that FtrABCD transports ferrous iron, 55Fe3+ and 55Fe2+ uptake assays were performed with strain 2308, the ftrA mutant and the ftrA reconstructed strain. For the 55Fe3+ uptake assay, sodium citrate was used to keep the ferric iron in a soluble form. For the 55Fe2+ uptake assay, ascorbate was used to reduce the Fe3+ to the Fe2+ form (Kammler et al., 1993; Robey and Cianciotto, 2002). As shown in Fig. 5A, the total amount of radiolabelled Fe2+ incorporated into the B. abortus ftrA mutant was significantly less than that incorporated into the parental 2308 strain and the reconstructed ftrA mutant over the 30 min time-course of the experiment. In contrast, all three of these strains incorporated the same amount of radiolabelled Fe3+ over this same period (Fig. 5B). This defect in Fe2+ incorporation by the B. abortus ftrA mutant could also be clearly observed when these data were used to calculate the rate of uptake of radiolabelled Fe2+ and Fe3+ as shown in Table 1. It is also important to note that when the proton motive force uncoupler carbonylcyanidep-trifluoromethoxyphenylhydrazone (FCCP) was included in parallel experiments to distinguish between energy independent binding of the radiolabelled iron to the bacterial cells and energy-dependent iron uptake, the time-dependent increase in incorporation of both radiolabelled ferrous and ferric iron was eliminated for all three strains (Fig. 5A and B). These experimental findings show that FtrABCD is an energy-dependent ferrous iron transporter in B. abortus 2308.
Table 1. Average rates of 55Fe3+ and 55Fe2+ uptake by B. abortus 2308, EAM003 (2308 ftrA) and EAM003RC (EAM003 ftrA+)
Average uptake rate pmole 55Fe min−1 109 cfu−1
These rates were calculated from the amount of energy-dependent radiolabelled iron incorporation measured for each strain at 1 and 30 min after the initiation of the experiment. All values represent the mean of three separate experiments ± standard deviation. See Experimental procedures for the specific details of this assay.
10.016 ± 1.22
5.105 ± 0.74
EAM003 (2308 ftrA)
9.515 ± 0.92
0.897 ± 0.12
EAM003RC (EAM003 ftrA+)
9.664 ± 1.36
5.054 ± 1.1
The expression of ftrA is responsive to iron-deprivation and low pH in B. abortus 2308
Both Northern blot analysis and RT-PCR analysis indicate that the ftrA, B, C and D genes are transcribed as an operon in B. abortus 2308, and all four of these genes display elevated levels of transcription in response to iron deprivation in a microarray analysis (data not shown). Using real-time RT-PCR analysis of RNA obtained from cultures grown under iron-deprived and iron-replete conditions, we verified the iron-responsive expression of ftrA (Fig. 6). The 2,3-dihydroxybenzoic acid biosynthesis gene dhbC gene was included as a positive control for the analysis shown in Fig. 6 because it has been shown by both Northern blot and gene fusion analyses that dhbC exhibits maximal expression in response to iron deprivation in B. abortus 2308 (Bellaire et al., 2003). To begin to define the cis-acting regulatory elements that control the iron-responsive expression of ftrA in B. abortus 2308, primer extension was employed to determine the transcriptional start site for this gene. A thymine residue (T) 72 nucleotides (nt) upstream of the predicted ftrA start codon was identified as the +1 start site for the ftrA transcript. Notably, computational analysis revealed a predicted binding site for the iron-responsive regulator Irr (also known as an ICE box) 55 nucleotides upstream of the ftrA transcriptional site (Fig. 1). Irr is a transcriptional regulator which often serves as an activator of iron acquisition genes in response to iron deprivation in the α-proteobacteria including Brucella (Small et al., 2009; Anderson et al., 2011). Correspondingly, data obtained from RT-PCR analysis showed that the elevated expression of ftrA observed in B. abortus 2308 in response to iron deprivation was abolished in an isogenic irr mutant, and was restored in a derivative of this mutant carrying a plasmid-borne copy of irr (Fig. 6). These results show that Irr is required for the iron-responsive expression of ftrA.
One of the distinctive features reported for the ftrABCD operon in Bordetella is that these genes appear to be independently responsive to both iron deprivation and low pH (Brickman and Armstrong, 2012), and as shown in Fig. 7 the B. abortus ftrA gene displays a similar pattern of expression. Specifically, the expression of ftrA in B. abortus 2308 was 24-fold higher at pH 5 than it was at pH 7 when this strain was grown in brucella broth (which is presumably an iron-replete medium) without the addition of a chelator.
A B. abortusftrA mutant exhibits significant attenuation in the mouse model
Brucella strains reside in acidified compartments during both early and late stages of their intracellular life cycle in host macrophages (Starr et al., 2012), an environment where Fe2+ potentially serves as an available iron source. Accordingly, since our experimental findings indicate that FtrABCD is the primary Fe2+ transporter in B. abortus 2308, it was important to evaluate the contribution of this transporter to virulence. As shown in Figs 8 and 9, the B. abortus ftrA mutant displayed significant attenuation in both cultured murine macrophages and experimentally infected C57BL/6 mice compared with the parental 2308 strain and a derivative of the ftrA mutant in which the ftrA locus had been reconstructed. These results clearly show that FtrABCD plays a critical role in the virulence of B. abortus 2308 in the mouse model. Moreover, they support the proposition that Fe2+ serves as an important iron source for Brucella strains in their mammalian hosts.
Several types of bacterial ferrous iron transporters have been discovered recently that rely on homologues of the well-characterized ferrous iron permease Ftr1p found in S. cerevesiae (Stearman et al., 1996) and other eukaryotic microbes (Askwith and Kaplan, 1997; Eichhorn et al., 2006; Larrondo et al., 2007; Ibrahim et al., 2010; Ziegler et al., 2011). A distinctive feature of the eukaryotic Ftr1p is that it forms a complex with the ferroxidase Fet3p, and this permease will not transport Fe2+ unless it is first oxidized to Fe3+ by Fet3p in a step that occurs concomitant with transport (Kwok et al., 2006). The best-characterized Ftr1-based bacterial ferrous iron transporters fall into three classes. The four component FtrABCD-type transporters found in Bordetella (Brickman and Armstrong, 2012) and Brucella (described in this report) are comprised of a periplasmic Fe2+-binding protein (FtrA), a periplasmic protein proposed to oxidize this Fe2+ to Fe3+ (FtrB), and two integral cytoplasmic membrane proteins, the Ftr1p-like iron permease FtrC and a ferredoxin that is thought to reoxidize FtrB and restore it Fe3+-oxidizing (e.g. ‘ferroxidase’) activity (FtrD). The three component EfeUOB-type transporter found in E. coli O157 strains (Cao et al., 2007) consists of a periplasmic protein thought to have both Fe2+-binding and ferroxidase activity, another periplasmic protein believed to reoxidize EfeO and restore its ferroxidase activity (EfeB), and the integral cytoplasmic membrane Ftr1p-like iron permease EfeU (Große et al., 2006; Cao et al., 2007; Rajasekaran et al., 2010).The third class of Ftr1-based bacterial ferrous iron transporters is exemplified by the FetMP-type transporters found in E. coli (Koch et al., 2011), Yersinia pestis (Fetherston et al., 2012) and Campylobacter jejuni (Chan et al., 2010). These transporters are made up of a periplasmic Fe2+ finding protein homologous to the FtrA protein known as FetM or P19, and the integral cytoplasmic membrane iron permease FetM. FetM is larger than its EfeU and FtrC counterparts, and unlike EfeU and FtrC, FetM is thought to directly transport Fe2+ (Koch et al., 2011) rather than requiring it first be oxidized to Fe3+ as has been proposed for EfeU (Rajasekaran et al., 2010) and FtrC (Brickman and Armstrong, 2012).
The results presented here not only demonstrate that FtrABCD is a ferrous iron transporter in B. abortus 2308, but they also show that this transporter plays an indispensable role in the virulence of this strain in the mouse model. The fact that the contribution of the Ftr transporter to virulence is not observed until some point after 1 week post infection may be a reflection of the cellular immune response against Brucella infections in mice being considerably stronger at 4 weeks post infection and later than it is at 1 week (Araya et al., 1989). It is well documented that IFN-γ increases the capacity of macrophages to deprive intracellular pathogens of iron through multiple mechanisms including downregulating the expression of transferrin receptors on the surface of these phagocytes (Byrd and Horwitz, 1989) and the increasing the activity of the iron efflux proteins Nramp (Cellier et al., 2007) and ferreportin (Nairz et al., 2008). Consequently, increased levels of IFN-γ at later times points post infection may result in the intracellular brucellae being exposed to more intense iron deprivation at 3 and 5 weeks post infection in mice than they are at 1 week post infection. But since little is known regarding how macrophage iron deprivation mechanisms influence the progression of Brucella infections, this proposition awaits experimental verification.
At least three different classes of Fe2+ transporters have been described in bacteria that are mammalian pathogens, the Feo-type systems (Cartron et al., 2006), the Yfe-/Sit-type systems (some of which also transport Mn2+) (Perry et al., 2012), and the FtrABCD, EfeUOB and FetMP systems that rely upon orthologues of the well-characterized eukaryotic iron permease Ftr1p for their activity (Rajasekaran et al., 2010; Koch et al., 2011). Feo-type transporters have been linked to virulence in Yersinia (Fetherston et al., 2012), Campylobacter (Naikare et al., 2006), Salmonella (Tsolis et al., 1996; Boyer et al., 2002), Porphyromonas (Dashper et al., 2005), Legionella (Robey and Cianciotto, 2002) and Helicobacter strains (Velayudhan et al., 2000); Yfe-/Sit-type transporters have been linked to virulence in Yersinia (Fetherston et al., 2012) and Shigella strains (Fisher et al., 2009); and a protein that transports Fe2+ across the outer membrane for subsequent transport into the cytoplasm presumably by an Feo-type transporter has been linked to virulence in Francisella tularensis (Ramakrishnan et al., 2012). The results presented demonstrate that an Ftr1-based Fe2+ transporter can also play an important role in bacterial virulence. This experimental finding takes on a special meaning for Brucella strains because it is well established that these bacteria reside in acidified compartments during their intracellular residence in macrophages (Celli et al., 2003; Starr et al., 2012). The vast majority of the iron present in the extracellular spaces of the mammalian host is Fe3+, due to the neutral pH and oxidizing nature of this environment (Anderson and Vulpe, 2009). In contrast, due to its reducing nature, the iron present within the intracellular environment of host cells is in a so-called ‘dynamic equilibrium’ between Fe3+ and Fe2+, and the ratio of these two forms of iron within specific intracellular compartments is dependent upon the pH of that compartment and the activity of host cell ferric reductases (which convert Fe3+ to Fe2+). Because of its propensity to react with reactive oxygen species, Fe2+ is potentially toxic, thus one of the main themes observed in cellular iron metabolism is that although Fe2+ is more soluble and biologically reactive than Fe3+, both prokaryotic and eukaryotic cells actively work to keep their intracellular levels of Fe2+ low (Kosman, 2010), and excess iron is stored and transported between cells as the less toxic Fe3+ form (Anderson and Vulpe, 2009). The consequence of this active conversion of Fe2+ to Fe3+ within host cells is that there are likely to be only a limited number of intracellular environments where Fe2+ would be expected to be readily available as an iron source for the brucellae. The so-called endolysosomal Brucella containing vacuoles (eBCVs) within which the brucellae reside during the early stages of their intracellular life cycle in host macrophages (Starr et al., 2012) potentially represent such an environment. Not only do these intracellular compartments have an acidic pH, where Fe2+ would be the predominant form, but they interact directly with the endolysosomal pathway, where Fe3+ is actively trafficked into the cell via transferrin mediated iron transport and subsequently converted into Fe2+ for intracellular distribution (Anderson and Vulpe, 2009). Thus, conceivably acidification of the BCVs during the early stages of infection not only provides an environmental stimulus for induction of the Brucella genes required for the maturation of these BCVs into replicative BCVs (Celli et al., 2003), but it may also play an important role in providing these bacteria with the iron they need to maintain their intracellular residence. Regardless of the precise intracellular environment(s) within which the brucellae encounter this divalent cation, the dramatic attenuation exhibited by the B. abortus ftrA mutant in mice strongly suggests that Fe2+ represents an indispensable iron source for these bacteria in their mammalian hosts.
Mice have been widely used as a model to evaluate the capacity of Brucella strains to establish and maintain chronic infections (Grillo et al., 2012). Thus, it is notable that the attenuation observed for the B. abortus ftrA mutant in cultured murine macrophages and mice stands in sharp contrast to the lack of attenuation observed for an isogenic dhbC mutant, which cannot produce siderophores, in this model (Bellaire et al., 1999). This apparent preference for Fe2+ over Fe3+ as a non-haem iron source in cultured mammalian macrophages and mice resembles the situation that has been described for Legionella pneumophila (Robey and Cianciotto, 2002; Allard et al., 2006; 2009), but differs from the results obtained with Salmonella enterica serovar Typhimurium (Nairz et al., 2009) and Mycobacterium tuberculosis (De Voss et al., 2000), which rely upon siderophores for iron acquisition during their intracellular replication in the host. These observations suggest that the siderophores produced by Brucella and Legionella strains, brucebactin (González Carreró et al., 2002) and legiobactin (Liles et al., 2000), respectively, may not be as well suited for capturing Fe3+ within the intracellular environment of host macrophages as those produced by Salmonella and Mycobacterium strains, e.g. salmochelin and mycobactin. But it is also possible that the activity of other iron transporters compensates for the lack of siderophore production in the B. abortus dhbC mutant in the mouse model, thereby masking the contribution of this siderophore in the in vivo setting as has been described in Shigella flexneri (Runyen-Janecky et al., 2003) and Y. pestis (Fetherston et al., 2012). In fact, Brucella strains produce at least four iron transport systems in addition to the one that utilizes the siderophore brucebactin (Roop et al., 2011), and it will be important to evaluate the virulence properties of mutants lacking different combinations of these transporters to accurately assess their individual and relative contributions to iron acquisition in the host.
As noted in the Introduction, no genes predicted to encode Feo- or Yfe/Sit-type Fe2+ transporters can be found in the available Brucella genome sequences. However, the ftrABCD locus appears to be highly conserved in Brucella strains (Roop et al., 2011). These observations coupled with the results reported here lead us to conclude that FtrABCD is the sole high-affinity Fe2+ transporter produced by these bacteria. One question that comes to mind is whether or not there are specific characteristics of this transporter that makes it uniquely suited for iron acquisition in the intracellular environments inhabited by Brucella strains in the host. One of the distinctive features of the eukaryotic Ftr1p permease is that Fe2+ is oxidized to Fe3+ before it is transported by this permease (Stearman et al., 1996; Kwok et al., 2006), and the recent observation that a Bordetella bronchiseptica mutant lacking the putative ferroxidase component (e.g. FtrB) of the FtrABCD transporter cannot transport Fe2+ suggests that the bacterial Ftr1p-like permeases have similar requirements (Brickman and Armstrong, 2012). It has been suggested that this oxidation of Fe3+ to Fe2+ serves to protect the Ftr1p permease from localized oxidative damage due to the toxicity of Fe2+ in oxidizing environments (Kosman, 2010). Previous studies have shown that preventing damage to cellular components resulting from chronic exposure of endogenous reactive oxygen species plays an important role in the virulence of Brucella strains (Steele et al., 2010). Thus, it is conceivable that the inherent ROS-resistance of FtrABCD allows it to serve as an effective iron transporter for Brucella strains during their intracellular residence in the host.
Haem represents an important iron source for Brucella strains in their mammalian hosts (Paulley et al., 2007). Haem transport across the outer membrane is mediated by BhuA, and the periplasmic binding protein-dependent ABC transporter comprised of the proteins designated BhuT, U and V shuttles haem across the cytoplasmic membrane (Ojeda, 2012). It was recently proposed that another bacterial Ftr1-based Fe2+ transporter, the EfeUOB system found in some E. coli strains (Cao et al., 2007), has the capacity to remove Fe2+ directly from the tetrapyrrole ring of haem in the periplasm and transport this divalent cation across the cytoplasmic membrane (Létoffé et al., 2009). But the putative ‘deferrochelatase’ activity of the EfeB component of this transporter has recently been called into question (Dailey et al., 2011). Nevertheless, the studies described in this report provide no evidence that FtrABCD plays a role in the capacity of Brucella strains to use haem as an iron source, since the B. abortus ftrA mutant can utilize this iron source with the same efficiency as the parental 2308 strain in an in vitro assay.
Studies to date suggest that the iron response regulator Irr is the predominant regulator of the iron metabolism genes in Brucella (Martínez et al., 2005; 2006; Anderson et al., 2011). So it is not surprising that this regulator is responsible for iron-responsive expression of the genes in the ftrABCD operon in B. abortus 2308. But another interesting, and potentially relevant feature of these genes, is that like their Bordetella counterparts (Brickman and Armstrong, 2012), the B. abortus ftr genes exhibit strong induction in response to exposure to acidic pH. Not only does a reduced pH environment favour the presence of soluble Fe2+ (Crichton, 2009), but studies have also shown that expression of the genes encoding the Ftr1-based transporter EfeUOB in E. coli is elevated in response to acid pH (Cao et al., 2007) and that the two-component regulator CpxAR is responsible for the acid-responsive expression of the efeUOB operon. Considering the documented importance of the acidification of the eBCVs for maintaining the intracellular life cycle of the brucellae (Starr et al., 2012), and the possibility that this reduced pH may play an important role in making Fe2+ a biologically relevant iron source for these strains, future studies will be directed at investigating the basis for the acid-responsive expression of the Brucella ftrABCD locus and identifying the genetic regulators responsible.
Bacterial strains and growth conditions
Brucella abortus 2308 and derivative strains were routinely grown on Schaedler agar containing 5% defibrinated bovine blood (SBA), in brucella broth, or in low-iron minimal medium (López-Goñi et al., 1992). For cloning experiments, E. coli strain DH5α was grown routinely on tryptic soy agar or in Luria–Bertani broth. Growth media were supplemented with kanamycin (45 μg ml−1) as necessary. To evaluate the growth characteristics of B. abortus 2308 and derivative strains at acidic pH under iron-restricted conditions, the strains were grown in brucella broth buffered at pH 6.0 or pH 7.0 with 100 mM of 2-(N-morpholino) ethanesulphonic acid (MES) or 3-(N-morpholino)propanesulphonic acid (MOPS), respectively, in the presence of 10 μM of the ferric-specific chelator, desferroxamine (DFO). Schaedler agar and brucella broth were obtained from Becton-Dickinson (Sparks, MD, USA), trypic soy agar and Luria–Bertani broth were obtained from EMD Chemicals (Gibbstown, NJ, USA), kanamycin sulphate was obtained from Genlantis (San Diego, CA, USA), DFO and MOPS were obtained from Sigma (St. Louis, MO, USA) and MES was obtained from Amresco (Solon, OH, USA).
Construction of a B. abortusftrA mutant
An isogenic mutant containing an in-frame deletion of the ftrA (BAB2_0840) gene was constructed from B. abortus 2308 using the non-polar, unmarked gene excision strategy described by Caswell et al. (2012). An approximately 1 kb fragment representing the upstream region of the gene to the 15th codon of the ftrA coding region was amplified by PCR using primers ftrA-Up-For 5′-GCGGATCCATGCAGCTTGTGGAGCCGCT′ and ftrA-Up-Rev 5′- GCTGAGCGGCACCATGAGC′. Similarly, a fragment containing the last two codons of the ftrA coding region to approximately 1 kb downstream region was amplified with primers ftrA-Down-For 5′-TATTGATCCATGGCGCGGC′ and ftrA-Down-Rev 5′-CGCTGCAGGAAGCATGACCTCGGTAATGCG′. The upstream fragment was digested with BamHI, while the downstream fragment was digested with PstI, and both fragments were treated with polynucleotide kinase in the presence of ATP. Both of the DNA fragments were included in a single ligation mix with BamHI/PstI-digested pNTPS138 (Spratt et al., 1986). The resulting plasmid (pAE001) was introduced into B. abortus 2308 by electroporation. The merodiploid transformants were obtained by selection on SBA + kanamycin. A single kanamycin-resistant clone was grown for ∼ 6 h in brucella broth, and then plated onto SBA containing 10% sucrose. Genomic DNA from sucrose-resistant, kanamycin-sensitive colonies was isolated and screened by PCR for loss of the ftrA gene, and an isogenic ftrA mutant derived from B. abortus 2308 was named EAM003. The ftrA mutation in this strain was verified by DNA sequence analysis and Southern hybridization.
Reconstruction of the parental ftrA in the B. abortusftrA null mutant
To confirm the link between the ftrA mutation and the phenotype exhibited by the B. abortus ftrA mutant, a derivative of this mutant was constructed in which the mutated ftrA allele was replaced by its wild-type counterpart. A 1704 bp fragment containing the intact ftrA gene and 509 bp upstream and 601 bp downstream of this gene was amplified from B. abortus 2308 genomic DNA with the primers ftrA-RC-For 5′-GCGGGCCCTTTCTACCACTACGACAT′ and ftrA-RC-Rev 5′-GCGCTAGCTAAGAAGCAGGAATGCCAG′. The DNA fragment was digested with ApaI and NheI and ligated with ApaI/NheI-digested pNTPS138. The resulting plasmid (pAE002) was introduced into the B. abortus ftrA mutant EAM003 by electroporation, and the selection of transformants containing the reconstructed ftrA locus and confirmation of their genotypes were performed as described above for the ftrA mutant. The B. abortus ftrA mutant, in which the ftrA locus was reconstructed, was named EAM003RC.
In vitro iron utilization assay
The capacity of the B. abortus strains to utilize different iron sources in vitro was investigated as described previously (Paulley et al., 2007) with some modifications. Brucella strains were grown in low-iron minimal medium at a starting inoculum of 106 cfu ml−1. Following incubation for 72 h at 37°C, bacterial suspensions in phosphate-buffered saline (pH 7.2) containing 109 cfu ml−1 were prepared. Aliquots of 100 μl of the bacterial cell suspensions were mixed with 4 ml 0.8% Noble agar. The mixtures were overlaid onto tryptic soy agar (TSA) plates containing 200 μM ethylenediaminedi (o-hydroxyphenylacetic) acid (EDDHA). In addition, the plates used for the Fe(NH4)2(SO4)2 utilization assay contained 2 mM sodium ascorbate in both the Noble agar overlay and the TSA plates. Seven-millimetre sterile filter paper (Whatman no. 3) disks were placed onto the plates, and 10 μl of a 50 mM solution of FeCl3, 50 mM Fe(NH4)2(SO4)2 or 20 mM haemin was added to the filter disks. Free iron was removed from the haemin stock solutions using a previously described procedure (Staggs and Perry, 1991). The plates were incubated for 72 h at 37°C with 5% CO2. Following this incubation period, the diameter (in millimetres) of the zone of bacterial growth around each disk was measured and recorded. FeCl3 and haemin were obtained from Sigma-Aldrich (St. Louis, MO, USA), Fe(NH4)2(SO4)2 was obtained from Acros Organics (Fair Lawn, NJ, USA), sodium ascorbate was obtained from MP Biomedical (Solon, OH, USA), EDDHA was obtained from Complete Green (El Segundo, CA, USA) and Noble agar was obtained from Becton-Dickinson (Sparks, MD, USA).
Determination of the transcriptional start site for ftrA by primer extension
The transcriptional start for the ftrA gene was determined by primer extension analysis performed on total RNA preparations obtained from B. abortus 2308 cultures grown for 72 h in low-iron minimal medium using the primer 5′-CAGCGGTTCTGAATAGATTTTTCAT′ and previously described methods (Robertson et al., 2000).
55Fe uptake assays
The methods described by Kammler et al. (1993) and Robey and Cianciotto (2002) were used to measure the capacity of B. abortus 2308, the ftrA mutant and the reconstructed mutant to transport 55Fe2+ and 55Fe3+ in an in vitro assay. Brucella strains were grown in low iron minimal medium to the late logarithmic phase at 37°C with shaking. Bacteria were then subcultured into low iron medium and allowed to grow again to the late logarithmic phase. The final cultures were harvested by centrifugation at 10 000 g for 5 min at 4°C, washed once in MOPS transport buffer (40 mM morpholinepropanesulphonic acid [MOPS-KOH (pH 7.0), 100 mM MgSO4, 0.5 mM CaCl and 0.2% glucose] and the bacterial pellet was resuspended in MOPS transport buffer to a cell density of 109 cfu ml−1 and kept on ice. Iron transport was initiated after 10 min warming at 37°C by addition of 55FeCl3 (Perkin-Elmer, Boston, MA, USA) to a final concentration of 3 μM in presence of 1 mM sodium citrate (for the Fe3+ iron uptake studies) or 5 mM sodium ascorbate (for the Fe2+ iron uptake studies). For the Fe3+ uptake assay, sodium citrate was added to the bacterial suspension in MOPS transport buffer to a final concentration of 1 mM. For the Fe2+ uptake assay, the 55FeCl3 stock was diluted in 100 mM sodium ascorbate and incubated for 30 min at room temperature to reduce the ferric iron before addition to the assay, and sodium ascorbate was also added to the bacterial suspension to achieve a final concentration of 5 mM. In parallel experiments, the proton motive force uncoupler carbonylcyanidep-trifluoromethoxyphenylhydrazone (FCCP) (Sigma, St. Louis, MO, USA) was added to bacterial suspensions in the transport buffer to a final concentration of 50 μM and incubated for 10 min before the addition of radiolabelled iron to distinguish between energy independent binding of the iron to the bacterial cells and energy-dependent uptake. At various times after the initiation of the uptake assay, 0.5 ml of the bacterial suspension was removed to a 0.45 μm nitrocellulose filter (Millipore, Bedford, MA, USA) and washed three times with 1 ml of 0.1 M LiCl at room temperature. The filters were then allowed to dry at room temperature for 30 min, and the amount of radioactive iron present was determined by scintillation counting.
Relative quantification of specific transcript levels using real-time RT-PCR
To quantify the relative ftrA transcript levels under low and high iron conditions, real-time RT-PCR was performed using total RNA isolated from cultures of B. abortus 2308, the irr mutant BEA2 and the complemented irr mutant BEA2.C (Anderson et al., 2011) following growth in low-iron minimal medium and low-iron minimal medium supplemented with 50 μM FeCl3 until late logarithmic phase. To evaluate the relative ftrA transcript levels under low and neutral pH conditions, the real-time RT-PCR assay was performed using total RNA isolated from cultures of B. abortus 2308 following growth in brucella broth buffered at pH 5.0 or pH 7.0 with 100 mM of 2-(N-morpholino) ethanesulphonic acid (MES) or 3-(N-morpholino)propanesulphonic acid (MOPS), respectively, until late logarithmic phase. In both cases, RNA was isolated from the bacterial cells using the procedures described by Caswell et al. (2012). Primers for 16S RNA (16S-RT Fwd 5′-TCTCACGACACGAGCTGACG-3′ and 16S-RT Rev 5′-CGCAGAACCTTACCAGCCCT′) were used as a control, while ftrA-specific primers (ftrA-RT Fwd 5′-GCATGTTGATTGGCCGAC′ and ftrA-RT Rev. 5′-GGTCGTTTCCTATGAGCT′) were used for evaluating relative ftrA mRNA levels. The dhbC-specific primers (dhbC Fwd 5′-GTGCCAAGCTTGGTCTGTACTTC′ and dhbC Rev 5′-CGTGGATTGTTTACCGGC′) were used for evaluating relative dhbC mRNA levels under low and high iron conditions. The relative abundance of transcripts was determined using the Pfaffl method (Pfaffl, 2001).
Virulence of the B. abortusftrA mutant in cultured murine macrophages and experimentally infected mice
The methods described by Gee et al. (2005) were used to evaluate the capacity of the B. abortus strains to survive and replicate intracellularly in cultured resident peritoneal macrophages obtained from C57BL/6 mice, and the methods described by Robertson and Roop (1999) were used to evaluate the capacity of these strains to establish and maintain chronic spleen infections in C57BL6 mice infected by the intraperitoneal route. The animal use protocols under which these experiments were performed were reviewed and approved by the East Carolina University Animal Care and Use Committee.
This work was supported by a grant from the National Institute of Allergy and Infectious Disease (AI68615) to R.M.R.