ABSTRACT
Mycobacterium tuberculosis requires iron for normal growth but faces a limitation of the metal ion due to its low solubility at biological pH and the withholding of iron by the mammalian host. The pathogen expresses the Fe3+-specific siderophores mycobactin and carboxymycobactin to chelate the metal ion from insoluble iron and the host proteins transferrin, lactoferrin, and ferritin. Siderophore-mediated iron uptake is essential for the survival of M. tuberculosis, as knockout mutants, which were defective in siderophore synthesis or uptake, failed to survive in low-iron medium and inside macrophages. But as excess iron is toxic due to its catalytic role in the generation of free radicals, regulation of iron uptake is necessary to maintain optimal levels of intracellular iron. The focus of this review is to present a comprehensive overview of iron homeostasis in M. tuberculosis that is discussed in the context of mycobactin biosynthesis, transport of iron across the mycobacterial cell envelope, and storage of excess iron. The clinical significance of the serum iron status and the expression of the iron-regulated protein HupB in tuberculosis (TB) patients is presented here, highlighting the potential of HupB as a marker, notably in extrapulmonary TB cases.
INTRODUCTION
Iron is an essential micronutrient for all aerobic bacteria, except lactobacilli and Borrelia burgdorferi (1). It plays an important role in vital biologic processes, including electron transport, where it participates in oxidation-reduction reactions by virtue of its transition between Fe3+and Fe2+ states. Iron, however, is harmful at high concentrations, as it mediates the formation of free radicals that damage macromolecules, like DNA and proteins. Fe2+, via the Fenton reaction, catalyzes the formation of hydroxyl radical (HO·), and the oxidized Fe3+ reacts with another molecule of hydrogen peroxide to form the hydroperoxyl radical (HOO·):
Despite the abundancy of iron, free iron is scarce at physiological pH, as it exists as insoluble iron oxides in the aerobic environment. Iron, in the form of the insoluble Fe(OH)2+ (solubility, 1.4 × 10−9 M at pH 7) is unavailable for bacteria that require 10−7 M iron for optimal growth (2). Pathogenic bacteria, including Mycobacterium tuberculosis, face an additional limitation of iron, as the mammalian host limits the amount of free iron by a process called nutritional immunity (3). Transferrin in the circulating plasma and lactoferrin present in extracellular fluids and polymorphonuclear leukocytes play important roles in reducing the availability of iron to the pathogen by virtue of their high affinity for Fe3+ (4, 5). Mycobacterium tuberculosis, like other mycobacteria, produces Fe3+-specific high-affinity low-molecular-mass (∼1,000 Da) compounds called siderophores for chelating the metal ion from insoluble and protein-bound iron. There are several reviews on mycobacterial siderophores and siderophore-mediated iron uptake mechanisms (5–9). Here, the focus is to present a comprehensive overview of iron homeostasis in M. tuberculosis and highlight its impact on the virulence of the pathogen. Recent advances on transcriptional regulation of siderophore biosynthesis, namely, the role of the iron-regulated histone-like protein HupB, revised models of transport of desferri- and ferrisiderophores, and the importance of storage of excess iron by the iron storage protein BfrB are presented here. Of clinical significance is the low serum iron status and the expression of the iron-regulated protein HupB in tuberculosis (TB) patients, reflecting the iron-limiting conditions faced by M. tuberculosis.
MYCOBACTERIAL SIDEROPHORES
Need for two types of siderophores.Mycobacteria produce two types of siderophores, the hydrophobic mycobactins and the water soluble carboxymycobactins that scavenge iron from the immediate environment; saprophytic mycobacteria produce exochelins as the predominant extracellular siderophore. Mycobactin is restricted to the cell envelope, which contains complex lipids, including the highly hydrophobic mycobacterium-specific mycolic acids. This lipid-rich organization renders the outer membrane of mycobacteria much more rigid than Gram-negative bacteria (10) and necessitates the presence of two siderophores for the uptake of iron. Unlike the TonB-dependent receptor-mediated internalization of the ferrisiderophore seen in Gram-negative organisms (11), it is highly likely that transfer of iron occurs from ferricarboxymycobactin from the outside to mycobactin localized close to the cytoplasmic membrane (8, 12). Transfer of iron from ferricarboxymycobactin to mycobactin has been demonstrated (13), and, as discussed later in this review, it is proposed that HupB, a 28-kDa iron-regulated cell wall-associated protein in M. tuberculosis (14), mediates this transfer.
Mycobactins.Mycobactins are essential for the in vivo growth and survival of M. tuberculosis (15). Almost all mycobacteria produce mycobactin under iron-limiting conditions, with the exception of M. paratuberculosis. In fact, Snow and White first identified mycobactin as a growth factor for the in vitro growth of this mycobacterial species (16). Snow and White, in their extensive structural elucidation of mycobactins produced by different species of this genus (17), identified the core mycobactin molecule with species-specific structural variation at residues R1 to R5 (Fig. 1a). The core nucleus consists of 2-hydroxyphenyloxazoline moiety linked by an amide bond to an acylated ε-N-hydroxylysine residue that is esterified at the α-carboxyl group with a β-hydroxy acid. The β-hydroxy acid is attached via an amide bond to the seven-membered lactam ring, formed by the cyclization of a second ε-N-hydroxylysine. The hydroxamic acid groups (N-OH) of the ε-N-hydroxylysines, the phenolate oxygen atom, and the nitrogen atom of the oxazoline moiety (circled in Fig. 1) chelate Fe3+ very effectively, explaining the high affinity of the molecule (∼1030) for the oxidized form of the metal ion (8), with poor binding seen with Fe2+.
Structural variations in mycobactin carboxymycobactin. The figure shows mycobactin T from M. tuberculosis, with the iron-chelating residues shown in red circles. R1 to R5 represent the residues that differ among mycobactins T, S, Av, M, H, and P from M. tuberculosis, M. smegmatis, M. avium, M. marinum, M. fortuitum, and M. phlei, respectively.
Figure 1 shows the substitutions at R1 to R5 contributing to the variations among the mycobactins produced by different mycobacterial species. R5, usually a long alkyl chain that contributes to the hydrophobicity of the molecule, differs among the various mycobactins in its length and unsaturation between α- and β-carbons. While this alkyl chain is predominantly seen at this position in most of the mycobactins, it is present at R3 in mycobactin from Mycobacterium marinum. At positions R1 and R2, a methyl group may or may not be present at the 6th position of the phenolic ring and the 5′ position of the oxazoline moiety. At R3 and R4, the variation can be seen in the alkyl substituents of the hydroxy acids. These differences contribute to the specificity of the mycobactin, making these molecules useful as taxonomical markers for the classification of mycobacteria. As pointed out by Snow (18), this variation possibly ensured that only the species that produced the molecule could effectively utilize it. The basis for this species specificity needs to be explored further, and it remains to be seen if this specificity is associated with the iron transport system.
Carboxymycobactins and exochelins.Carboxymycobactins are expressed as the sole extracellular siderophore by pathogenic mycobacteria, including M. tuberculosis (19) and M. bovis strains (20). These molecules have the same chemical structure of the mycobactins produced by the respective strain but carry a shorter acyl chain at R5; in fact, the carboxymycobactins upon high-performance liquid chromatography (HPLC) purification are not seen as a single species but are a heterogenous group of molecules that differ in the length of this acyl chain (20). They, like mycobactin, have high affinity for Fe3+ and can remove insoluble and protein-bound iron. They are the sole extracellular siderophores in pathogenic mycobacteria and, as will be discussed later, experimental evidence shows that the disruption of their biosynthesis affects the growth and viability of these organisms. It is unclear why the saprophytic Mycobacterium smegmatis, utilizing the peptidic exochelins as the major extracellular siderophore, produces these compounds, albeit in low concentrations (21). The two extracellular siderophores were first differentiated by virtue of the extractability of their ferric forms into chloroform (22); the carboxymycobactins partitioned into the chloroform layer, and the exochelins were restricted to the aqueous layer.
Exochelins, as known to date, are produced by free-living bacteria only. They are peptides, and structural elucidation of the exochelins from M. smegmatis (23) and exochelin MN from Mycobacterium neoaurum (24) show them to be pentapeptides and hexapeptides, respectively, that strongly chelate Fe3+ via the hydroxamic acid residues generated as a result of modifications of the amino acids ornithine and histidine. The genome of M. tuberculosis does not show the presence of any of the genes associated with the biosynthesis and transport of exochelin (25). Iron acquisition by the human pathogen Mycobacterium leprae is a mystery, as it lacks both the mbt biosynthetic machinery and the exochelin-linked genes (26) Interestingly, the pathogen acquired iron from ferriexochelin MN could not take up iron from ferriexochelin MS and other mycobacterial carboxymycobactins (27).
SIDEROPHORE BIOSYNTHESIS IN M. TUBERCULOSIS
In M. tuberculosis, the genes encoding the proteins for the assembly of the mycobactin carboxymycobactin are organized in two clusters, namely, the mbt-1 (28) and mbt-2 (29) loci. The 24-kb mbt-1 locus consists of 10 genes, mbtA to mbtJ, and contains the information for synthesizing the core structure of the mycobactin molecule (Fig. 2a). The mbt-2 cluster, composed of four genes, mbtK to mbtN (Fig. 2c), incorporates the hydrophobic aliphatic side chain onto the mycobactin backbone (29). A recent study established the functionality of eight of the genes in the mbt-1 cluster by using a systematic mutational approach in M. smegmatis (30). Figure 2b gives a schematic representation of the nonribosomal biosynthetic pathway for the assembly of the core mycobactin molecule. This involves the synthesis of salicylate by MbtI, hydroxylation of lysine by MbtG giving N6-hydroxy Lys, and assembly of the mycobactin backbone by the megasynthase complex, consisting of three nonribosomal peptide synthetases (NRPS; MbtB, MbtE, and MbtF) and two polyketide synthases (PKS; MbtC and MbtD). The roles of MbtJ and MbtH are yet to be identified. The core mycobactin is then acylated by a long-chain fatty acyl group to form mycobactin, with the reactions mediated by the products of the genes in the mbt-2 locus, namely, MbtL (FadD14, strain Rv1344), MbtM (FadD33, strain Rv1345), MbtN (FadE14, strain Rv1346), and MbtK (lysine N-acetyltransferase, strain Rv1347c). After the formation of the acyl chain by MbtL, MbtM, and MbtN (Fig. 2c), it is transferred (29) to the core mycobactin by the MbtK-fatty acyl complex (Fig. 2d), forming the functional mycobactin. The mbt-2 cluster has been characterized based on in vitro studies, and the roles of these genes in siderophore biosynthesis are yet to be supported by genetic evidence. Mycobactins from different mycobacterial species differ in the length of this acyl chain, and structural analysis of the proteins encoded by the mbt-2 cluster in different mycobacterial species may explain the heterogeneity of the aliphatic side chain added by this locus. For example, MbtK mediating the acetylation of lysine residues was found to show amino acid variations in the corresponding enzyme from Nocardia, a genus closely linked to mycobacteria (29) that produces mycobactin-like siderophores called nocobactins under low-iron conditions. Whether these changes in MbtK are associated with the variations in the acetylation of lysine residues in nocobactin needs to be experimentally proved.
Biosynthesis of mycobactin. (a) Organization of the mbt genes in the mbt-1 locus and the locations of the two IdeR boxes. (b) Sequential steps involved in the nonribosomal synthesis of the core mycobactin molecule. The figure shows the different Mbt enzymes and ATP required to drive the specific reaction(s). ε-RHN-lysine represents an active intermediate of lysine that interacts with MbtE to form ε-RHN-lysine-MbtE. (c and d) Genes in the mbt-2 locus and the reactions leading to the formation of the MbtK-acyl complex (c), which transfers the acyl moiety to the core mycobactin molecule (d).
TRANSCRIPTIONAL REGULATION OF MYCOBACTIN BIOSYNTHESIS: ROLE OF IdeR AND HupB
IdeR.Iron-dependent regulator (IdeR), first described by Schmitt and coworkers (31), represses mycobactin biosynthesis in the presence of iron. Structural elucidation of this 230-amino-acid-long protein coupled with functional characterization (32–35) explains its role as a transcriptional regulator. It has three domains, with domain 1 (amino acids 1 to 73) binding DNA by virtue of its helix-turn-helix motif, domain 2 bearing the two metal-binding sites constituting the dimerization domain (residues 74 to 140), and domain 3, called the SH3-like domain (amino acids [aa] 151 to 230). In the absence of iron, apo-IdeR is loosely held as a dimeric species that cannot bind DNA, and effective dimerization occurs only upon the addition of a divalent metal ion (35). In the absence of metal ions, the SH3 domain prevents the formation of the stable dimer by binding to residues 125 to 139, called the tether region. When the metal binding sites are occupied, the SH3 domain moves away from the tether region, and the stable IdeR-Fe2+complex, with its four iron atoms, can bind DNA strongly. Although iron is the preferred metal ion, IdeR can bind other divalent metal ions, such as Mn2+, Zn2+, Co2+, Ni2+, and Mg2+, but at higher concentrations.
IdeR, under iron-sufficient conditions, functions predominantly as a negative regulator, switching off the synthesis of genes associated with iron acquisition, but acts as a positive regulator of the iron storage genes bfrA and bfrB (36). The earlier report on the essentiality of IdeR in M. tuberculosis (37) was substantiated by a recent study in which the role of IdeR on iron homeostasis and virulence was established using a conditional ideR mutant of M. tuberculosis that failed to survive in macrophages and experimental mice (38). The mutant strain showed high levels of iron due to uncontrolled mycobactin synthesis and low levels of the storage proteins BfrA and BfrB.
HupB (Rv2986c).HupB is annotated as a 22-kDa DNA-binding histone-like protein (Hlp) in the genome of M. tuberculosis. The protein, containing 214 amino acids, has an N-terminal region of 90 amino acids homologous to the Escherichia coli histone-like DNA-binding HU class of nucleoid proteins and a highly basic C-terminal region rich in lysine and arginine that is unique to mycobacteria. The presence of these basic amino acids gives the protein a high pI value of 12.5 and accounts for its altered electrophoretic mobility as a 28-kDa protein against its calculated molecular mass of 22 kDa. Mycobacterial HupB shows considerable sequence variations in the C-terminal region. Phylogenetic analysis of the protein from different mycobacterial species grouped the pathogenic and nonpathogenic members as well-separated clusters (see Fig. S1 in the supplemental material). The heterogeneity of HupB was also seen among the members of the highly conserved M. tuberculosis complex, the significance of which remains to be understood. Notable was the deletion of a 27-bp stretch, encoding the 9 amino acids at positions 137 to 145 in M. bovis BCG Pasteur. Also, M. marinum, a member of the M. tuberculosis complex, forms a separate cluster with Mycobacterium ulcerans (Fig. S1).
HupB, which has also been named HLPMt (histone-like protein in M. tuberculosis), mycobacterial DNA-binding protein 1 (MDP1), and laminin-binding protein (LBP) have been implicated in several biological functions, including immunoproliferation (39), adhesion (40), assembly of the cell wall (41), and recombination (42). The association of HupB with iron metabolism in M. tuberculosis was first demonstrated by Yeruva and coworkers, who identified the protein as a 28-kDa cell wall-associated protein in organisms (14) grown in iron-limiting medium (0.02 μg Fe ml−1; 0.36 μM Fe). Under high-iron conditions, hupB transcription is repressed by the IdeR-Fe2+complex, as seen from its interaction with the two IdeR boxes located at positions −5 and −127 upstream of hupB (43).
The functional characterization of HupB in iron homeostasis was made possible by the generation of a hupB knockout (KO) mutant strain of M. tuberculosis (44). In iron-limiting medium, the KO strain, unlike the wild-type (WT) organisms, expressed markedly low levels of mycobactin and carboxymycobactin that were restored upon hupB complementation, indicating the role of HupB in promoting mycobactin biosynthesis. It is therefore not surprising that the hupB KO mutant strain, like the mbtB KO mutant strain of M. tuberculosis (15), could not multiply inside macrophages, as both of them were unable to produce mycobactin.
How do the levels of iron, IdeR, and HupB regulate the mbt biosynthetic machinery?IdeR in the presence of iron negatively regulates the mycobacterial mbt biosynthetic machinery by the classical repression mechanism. The IdeR-Fe2+ complex (36) binds specifically to a 19-bp consensus sequence called the iron box or IdeR box (5′-TTAGGTTAGGCTAACCTAA-3′) in the promoter DNA of the mbt genes. This blocks the transcription of the mbt genes by RNA polymerase (Fig. 3a) when the intracellular iron is high. Under iron-limiting conditions, as a stable IdeR-Fe2+ complex is not formed, the IdeR box remains unoccupied and enables RNA polymerase to transcribe the mbt genes (36). However, the mere absence of the IdeR-Fe2+ complex at the IdeR box is not sufficient to initiate transcription and requires the binding of HupB to a 10-bp sequence (5′-CACTAAAATT-3′) called the HupB box, located immediately upstream of the IdeR box (44). This explains the restoration of mycobactin biosynthesis upon complementation of the hup KO mutant strain with hupB. The presence of a functional HupB box in the hupB promoter DNA showed that HupB potentiated not only mycobactin production but also its own synthesis.
Transcriptional regulation of mycobactin biosynthesis: a schematic model. (a) Repression of the mbt genes and hupB by IdeR-Fe2+ complex. In the presence of iron, IdeR forms a stable dimeric IdeR-Fe2+ complex, with two iron atoms in each of the monomeric units. This complex binds to the IdeR box iron box in the promoter DNA of the mbt genes and blocks their transcription. IdeR-Fe2+ complex also binds to the two IdeR boxes in the hupB promoter region (positions −5 and −127), repressing the transcription of hupB. (b) Sequence of events under iron-limiting conditions leading to expression of mycobactin. When the IdeR-Fe2+ complex cannot be formed under iron-limiting conditions, HupB binds to the HupB box, located upstream of the IdeR box, and promotes the transcription of the mbt genes. Further, it positively regulates its own synthesis by binding to the HupB box in its promoter DNA.
Occupancy of the IdeR box or the HupB box by the respective proteins will determine if the mbt machinery will be repressed or activated. This outcome is dictated by the iron concentration. While the intracellular iron concentration controlling mycobactin biosynthesis in vivo is not known, organisms grown in axenic medium can be grown in defined low- and high-iron medium containing a calculated amount of iron. Thus, maximal mycobactin and carboxymycobactin were produced by M. tuberculosis in low-iron medium containing 0.36 μM Fe (14), with negligible levels of both the siderophores in high-iron medium (144 μM Fe). It is possible to maintain such defined iron-limiting conditions in vitro that allow the cells to continue synthesizing the siderophores. In fact, mycobactin levels in M. smegmatis can reach values as high as 10% of the cell dry weight (8).
The activation of the mbt machinery is proposed to occur when HupB is positioned in the HupB box in the promoter DNA of the mbt genes. In the mobility shift assays, a surprising and unexpected finding was the requirement of iron for the binding of HupB to the HupB box in the mbtB promoter DNA. This may sound paradoxical considering (i) there is a negative correlation of HupB expression with iron levels and (ii) if iron is present, the IdeR-Fe2+ complex will be formed, and it will occupy the IdeR box and prevent HupB from occupying the adjacent HupB box. This was addressed by determining the concentration of iron needed for IdeR and HupB to bind their respective binding regions in the mbtB promoter DNA. Mobility shift studies revealed IdeR required severalfold-higher concentrations of iron (≥200 μM Fe) to bind the mbtB promoter, while HupB needed at ≤25 μM Fe. Thus, at 200 μM Fe, IdeR functions as a repressor and downregulates both mycobactin and HupB expression. Neither HupB nor the two siderophores were detected in wild-type M. tuberculosis grown in medium with ≥200 μM Fe (44). When iron levels were lowered, HupB was induced much earlier than the siderophores and was detected even at 144 μM Fe in the growth medium. It will promote the transcription of mbtB by occupying the HupB box in the mbtB promoter DNA, a process that can occur when the IdeR box is empty. This situation is possible only when the iron concentration drops to a value that does not allow the formation of the IdeR-Fe2+ complex. This explains the low mycobactin production by the hupB-complemented KO strain grown in high-iron medium; despite the constitutive expression of HupB, there was no mycobactin production in the presence of iron, as the IdeR-Fe2+ complex possibly occupied the IdeR box and blocked the binding of HupB to the HupB box (44). When low-iron conditions prevail, HupB will not only maintain its own levels but will ensure sufficient mycobactin production to scavenge iron from the immediate environment.
While such defined conditions are possible in vitro, it is not appropriate to extrapolate them to in vivo situations, where fluctuations in iron levels will occur without reaching the so-called high- and low-iron conditions established in vitro. This is because subtle changes will be perceived by the pathogen that will switch the mbt machinery on or off. Thus, when the intracellular iron level goes below the concentration needed for formation of the IdeR-Fe2+ complex, the mbt machinery will be activated, and when sufficient iron is taken up, there will be repression of mycobactin synthesis. The mbt biosynthetic machinery is likely to be switched on or off over a narrow range of iron concentrations without reaching the so-called high- and low-iron conditions achieved in vitro.
Other proteins influencing mycobactin biosynthesis include the MmpS4 and MmpS5 proteins (45), although the exact mechanism of action remains to be understood.
TRANSPORT OF IRON ACROSS THE MYCOBACTERIAL MEMBRANE
Early studies by Ratledge and Dover (8, 46) and recent reports (45, 47–50) have addressed the transport of iron across the mycobacterial envelope and export of the desferrisiderophore from the cytoplasm in M. smegmatis and M. tuberculosis. Unlike the wealth of information on the ferrisiderophore receptors in E. coli (11), identification and characterization of iron-regulated envelope proteins (IREPs) as iron transporters have been slow to come in mycobacteria, mainly due to the difficulties faced in the genetic manipulation of mycobacteria. To date, the ferriexochelin receptor in M. smegmatis and the IrtAB cytoplasmic transporter in M. tuberculosis have been characterized as transporters of iron, but a complete understanding of the transport mechanism remains elusive.
A 29-kDa ferriexochelin receptor in M. smegmatis.Among several IREPs expressed by M. smegmatis, a 29-kDa protein was characterized as a receptor for ferriexochelin (51, 52), both by demonstrating direct interaction of the purified protein with ferri-exochelin MS and by inhibiting ferriexochelin-mediated iron uptake by preincubating live organisms with specific antibodies against the 29-kDa IREP. Several mycobacteria express this protein (53), and it was also identified in an in vivo-derived Mycobacterium avium strain isolated from the liver of experimentally infected C57 black mice (53). Another IREP is the 21-kDa protein in M. neoaurum that was coexpressed with mycobactin and exochelin (54). Interestingly, an identical 21-kDa band was seen in the cell envelope fractions of M. leprae isolated from armadillo liver (53). It is possible that this IREP mediates iron uptake, as this human pathogen can take up iron only from exochelin MN from M. neoaurum (27) and not from other mycobacterial siderophores.
IrtAB ABC transporter in M. tuberculosis.In M. tuberculosis, several researchers (55–57) reported iron-regulated proteins whose specific roles in iron acquisition are suggestive and not established. The first protein shown to play a definitive role in this pathogen is IrtAB, first documented in 2006 (58). IrtAB is formed by the association of the membrane proteins IrtA and IrtB, encoded by irtA and irtB (strains Rv1348 and Rv1349), respectively. There is 34% identity among these two proteins in the transmembrane and carboxy-terminal domains, and they differ in the N-terminal region of a stretch of 272 amino acids present only in IrtA. Using an irtAB KO mutant strain of M. tuberculosis, the protein was functionally characterized as an iron transporter that mediated the internalization of iron using ferricarboxymycobactin as the source of iron. In addition, the N-terminal domain of IrtA, referred to as IrtA-NTD, was shown to bind flavin adenine dinucleotide (FAD) and hypothesized to function as an FAD-dependent reductase (47), thereby implicating the IrtAB system in not only transporting the metal ion but also catalyzing its reduction to Fe2+. This irtAB KO mutant strain failed to survive inside human macrophages and in experimental mice, highlighting the in vivo significance of this transporter.
MmpL4, MmpS4, MmpL5, and MmpS5 proteins.MmpL4, MmpL5, and the associated MmpS4 and MmpS5 proteins were first shown to play crucial roles in iron acquisition and virulence of M. tuberculosis by Wells and his group (45). They developed specific knockout mutants and showed that MmpS4 and MmpS5, with their respective MmpL4 and MmpL5 transport proteins, formed a novel siderophore export system for mycobactin and carboxymycobactin. Interestingly, in a subsequent study (59), it was demonstrated that the addition of exogenous siderophore to the ΔmmpS4 ΔmmpS5 double KO mutant inhibited its growth due to the toxicity of the accumulated desferrisiderophore. While the mutant strain was able to take up the ferric forms of carboxymycobactin and mycobactin and utilize the iron, it was unable to export the desferric form that clearly implicated that recycling of the siderophore was taking place in M. tuberculosis. However, the picture is far from complete, as the outer membrane exporter has yet to be identified.
In this context, it is appropriate to mention that the transcriptome data (GEO accession no. GSE53254) of the hupB knockout mutant (44) showed a 3-fold lower transcript level of both mmpS5 and mmpL5, with an ∼1.5-fold decrease in the mmpS4 and mmpL4 transcripts. Experimental evidence is necessary to establish if HupB, by regulating the expression of the MmpL4, MmpL5, MmpL4, and MmpS4 proteins, plays an indirect role in siderophore export (44).
Proposed model for the transport of iron in M. tuberculosis.Figure 4 is a diagrammatic representation of the proposed model of iron uptake in M. tuberculosis. When faced with low levels of free iron, the pathogen releases desferricarboxymycobactin into the immediate environment. The MmpS4 and MmpS5 and the associated MmpL4 and MmpL5 transport proteins are associated with this export process (45), and, once released the desferricarboxymycobactin, chelates Fe3+ from insoluble or protein-bound iron and forms ferricarboxymycobactin. Here, it is proposed that iron from ferricarboxymycobactin is transferred to mycobactin in the cell envelope of M. tuberculosis, with HupB functioning as the iron transporter. HupB is proposed as the iron transporter based on its surface localization (60, 61), its property to bind Fe3+, and interaction with ferricarboxymycobactin and ferrimycobactin (62). HupB is not only seen in the 50S ribosomal subunit (63) but in the cell wall (14), specifically on the cell surface (60). The purified protein bound radiolabeled ferricarboxymycobactin and ferrimycobactin in a dose-dependent manner, the specificity of which was demonstrated by displacement of the bound label upon the addition of cold ferrisiderophore (61). As HupB binds Fe3+ (62), it is hypothesized that it mediates the transfer of the metal from ferricarboxymycobactin to mycobactin, also reported in another study (13). The IrtAB transporter, localized on the cytoplasmic membrane, mediates the internalization of the iron (47) after reduction of Fe3+ to Fe2+ by the FAD-dependent reductase activity of the IrtA protein. The metal ion, as Fe2+, is transported into the cytoplasm by an energy-dependent process, and once inside, it is utilized for various metabolic processes, and the excess iron is stored in BfrA and BfrB.
Proposed model for the transport of iron in M. tuberculosis. Desferricarboxymycobactin (CMb) produced by M. tuberculosis upon iron limitation is exported to the outside, a process facilitated by the MmpL4 and MmpS4, and MmpL5 and MmpS5 proteins, as proposed by Wells et al. (45). It chelates Fe3+ present as insoluble or protein-bound iron and forms ferricarboxymycobactin. The surface-exposed HupB mediates the transfer of Fe3+ from ferricarboxymycobactin to mycobactin (Mb) located on the cytoplasmic membrane. The role of another protein participating in this transfer is a possibility (represented by a question mark) that needs to be explored. Iron from the mycobactin is acted upon by the IrtAB iron transporter that catalyzes the reduction of Fe3+ to Fe2+ and mediates the ATP-dependent transfer of iron into the cytoplasm, where it is utilized for various metabolic processes; the excess iron is stored in BfrA and BfrB.
The ESX-3 secretion pathway has been implicated in iron transport (48–50), and experimental evidence shows the essentiality of this pathway for the survival of M. tuberculosis inside macrophages. The transcription of esx-3 is controlled both by iron and zinc levels, with the participation of the respective iron and zinc regulators IdeR and Zur. In M. smegmatis, only iron influenced the expression of the components of esx-3. While knockout mutants show altered iron uptake, the exact role of the ESX-3 on iron transport is unclear.
Microvesicles (MV), containing mycobactin (and not carboxymycobactin), were implicated in iron transport in M. tuberculosis (64). The study showed that that these MVs supported the growth of a mycobactin-defective mutant strain in low-iron medium. If MV-associated mycobactin scavenged the iron, the need to secrete carboxymycobactin does not arise. The role of mycobactin as a carrier for iron needs to be explored further, as it is not clear as to how mycobactin, contained within MVs, can take up iron from the environment and deliver the metal ion to the organism.
TOXICITY OF IRON: PROTECTIVE ROLE OF THE IRON STORAGE PROTEINS BfrA AND BfrB IN M. TUBERCULOSIS
Considering the toxicity of iron at higher concentrations, it is not surprising that all living organisms limit the amount of free iron by storing the excess as protein-bound iron. In bacteria, ferritins and or the heme-containing bacterioferritins serve as iron storage proteins. Mycobacterial bacterioferritin was first reported in M. paratuberculosis (65) and later in M. leprae (66). Genome analysis of M. tuberculosis (25) identified bfrA (Rv1876), encoding the bacterioferritin BfrA, and bfrB (Rv3841), encoding the ferritin-like protein BfrB. These two proteins, with molecular masses of 18 and 20 kDa, respectively, aggregate to form macromolecular structures consisting of 24 subunits that can hold 600 to 2,400 iron atoms per molecule.
IdeR, in the presence of iron, positively regulates the expression of both iron storage genes. In vitro studies showed that IdeR in the presence of iron bound the promoter DNA of bfrA and bfrB (36). In-depth analysis of the transcriptional regulation of bfrB showed that IdeR exerted its effect by countering the repressor effect of Lsr2, another histone-like DNA-binding protein (67). It was proposed that Lsr2 functioned as a repressor of bfrB transcription under low-iron conditions, but in the presence of iron, Lsr2 was displaced by IdeR-Fe2+ complexes that bound four tandemly placed IdeR boxes in the bfrB promoter, thereby promoting the transcription of bfrB.
Studies with the bfrA and bfrB knockout mutants established the importance of the iron storage proteins (38, 68, 69), particularly BfrB, whose absence resulted in iron toxicity. The bfrB mutants did not survive in experimental mice, showing that BrfB was essential for the virulence of M. tuberculosis. In the absence of this protein, iron was not sequestered, resulting in oxidative stress. As mentioned earlier, similar iron toxicity was observed in the ideR mutant strain of M. tuberculosis that produced severalfold-higher levels of mycobactin due to the complete loss of regulation by IdeR on the mbt biosynthetic biochemistry and the failure to upregulate bfrB.
IRON AND TUBERCULOSIS
Iron and virulence of M. tuberculosis.Reviews on iron and TB (8, 70) have addressed host iron levels and susceptibility to TB; reports of studies in experimental animals have shown disease development upon administration of iron that overrides the response of the mammalian host to limit iron in an infection. Iron is essential for the normal growth of M. tuberculosis inside the macrophages, where the iron levels are low, ranging between 1 and 10 ng ml (8). This is maintained in spite of the high flux of the metal ion inside the macrophages due to the destruction of erythrocytes and internalization of iron via specific cell surface receptors for transferrin, lactoferrin, and hemoglobin-haptoglobin. However, most of the iron is transferred to the bone marrow, and any free iron is bound by transferrin and lactoferrin. Lactoferrin, by virtue of its ability to hold the metal ion even at acidic pH, plays an important role in withholding iron to M. tuberculosis residing within the alveolar macrophages in patients with pulmonary TB. Thus, the elaboration of the siderophore machinery is necessary, as shown by the sequestration of the metal ion by M. tuberculosis from holotransferrin (71, 72) and from hololactoferrin (73, 74).
If iron was made available to the pathogen, as done experimentally in macrophage cultures or in experimental animals infected with M. tuberculosis, there was enhanced multiplication of the pathogen (3, 5, 75–77). When the iron was given along with iron chelator deferoxamine or apo-transferrin, there was inhibition of growth of the pathogen, clearly establishing the role of iron in TB. These findings were later substantiated by using KO mutants with specific defects in iron acquisition. As discussed earlier, mycobactin biosynthesis and transport via the siderophore system are essential for the in vivo survival of the pathogen, reaffirming the association of iron acquisition with virulence of the pathogen.
In vivo expression of HupB: can HupB serve as a biomarker in TB patients?That the pathogen faces iron deprivation within macrophages is evident from the upregulation of the mbt genes (78, 79). It is thus likely that one or more components of the iron acquisition machinery can serve as marker(s) to reflect the iron status of the pathogen. Sivakolundu et al. (80) and Sritharan et al. (81) identified HupB as a putative marker in two independent studies conducted on TB patients in India (80, 81). In both of the studies performed on pulmonary and extrapulmonary TB patients, two important observations were made. First, there was negative correlation of the titer of anti-HupB antibodies with serum iron levels, and second, anti-HupB antibody titers in extrapulmonary TB patients were notably high, with levels exceeding those seen in pulmonary TB patients. When the full-length HupB was used as antigen (80), anti-HupB antibodies in the serum level of extrapulmonary TB patients (optical density at 450 nm [OD450], 1.230 ± 0.341, P < 0.05, compared to 0.230 ± 0.042 in healthy endemic controls) were higher than those seen in smear-positive pulmonary TB patients (OD450, 0.678 ± 0.205; P < 0.05). Interestingly, the titers were low in the household contacts of these patients (mean ± standard deviation [SD], 0.313 ± 0.128), an observation of relevance due to the endemicity of the disease. As mentioned, the antibody titer showed a statistically significant (P < 0.01) negative correlation with serum iron and total iron binding capacity (TIBC); for example, the circulating iron level in pulmonary TB patients was only 47.70 ± 39.48 μg dl−1, compared to 107.74 ± 45.74 μg dl−1 in the endemic healthy controls (P < 0.05). The serum ferritin levels in TB patients were high, with some pulmonary patients showing values as high as 2,500 ng ml−1 compared to the mean ± SD of 89.07 ± 141.50 ng ml−1 in the healthy controls. When three different antigenic fragments of HupB were used as antigens (81), HupB-F2 antigen bearing amino acids 63 to 161 was highly promising as an antigen, detecting remarkably high levels of anti-HupB antibodies in the serum of extrapulmonary TB patients. Also, the high antibody titers in pulmonary TB patients with relapse of the disease suggested prolonged exposure of the tubercle bacilli, possibly as dormant bacilli to iron-limiting conditions inside the human host.
The potential of HupB as a marker for TB stems from the finding that it can identify TB patients in a region that is endemic for the disease and where BCG is used as a vaccine. This will have considerable impact on the diagnosis of extrapulmonary TB, which is often difficult to diagnose with conventional tests. In fact, using hupB-specific primers, PCR identified M. tuberculosis as the causative organism for the chronic inflammatory disease Takayasu's arteritis (82), comparable to the reference control amplified using IS6110-specific primers.
CONCLUDING REMARKS
It is evident that iron is essential for the growth and survival of M. tuberculosis. The pathogen expresses the siderophores mycobactin and carboxymycobactin under iron-limiting conditions. It does not possess the exochelin machinery elaborated by the nonpathogenic M. smegmatis and is thus dependent on the carboxymycobactin and mycobactin uptake system for acquiring this essential micronutrient. Thus, any disruption of the mbt biosynthetic machinery affects its growth and survival, as demonstrated experimentally with the mbtB and hupB KO mutant strains. HupB plays an important role in sensing iron levels and functions as a positive regulator of mycobactin biosynthesis. The clinical significance of the protein can be inferred from the high titers of anti-HupB antibodies in the serum of TB patients, the majority of whom presented with low serum iron levels.
Remarkable progress has been made in understanding of the transcriptional regulation of the mycobactin biosynthetic machinery and expression of the iron storage proteins BfrA and BfrB. Iron homeostasis is a tightly controlled process, balancing iron uptake, utilization, and storage. Considering the essentiality of iron and the various regulatory controls used by the pathogen to maintain optimal iron for its growth and survival, it will be worth exploring these pathways to identify potential drug, vaccine, and diagnostic targets. The diagnostic potential of HupB, particularly for extrapulmonary TB, is worth exploring. There is a need for better control measures for this dreaded disease, and advances in diagnosis, development of novel drugs, and vaccines are the need of the hour, with the iron acquisition machinery offering ample scope for this unmet need.
ACKNOWLEDGMENTS
I thank the funding bodies CSIR, DBT, and DST (Government of India) for funding some of the work presented in this review. I also thank the Commonwealth Commission for supporting the faculty fellowship and Split-Site student fellowships for my doctoral students to work in the Tuberculosis Research Group at the Veterinary Lab Agency, United Kingdom.
FOOTNOTES
- Accepted manuscript posted online 11 July 2016.
- Address correspondence to srimanju{at}yahoo.com.
Citation Sritharan M. 2016. Iron homeostasis in Mycobacterium tuberculosis: mechanistic insights into siderophore-mediated iron uptake. J Bacteriol 198:2399–2409. doi:10.1128/JB.00359-16.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00359-16.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- Copyright © 2016, American Society for Microbiology. All Rights Reserved.