Ferric iron reductases and their contribution to unicellular ferrous iron uptake

Abstract

Iron is a necessary element for nearly all forms of life, and the ability to acquire this trace nutrient has been identified as a key virulence factor for the establishment of infection by unicellular pathogens. In the presence of O₂, iron typically exists in the ferric (Fe³⁺) oxidation state, which is highly unstable in aqueous conditions, necessitating its sequestration into cofactors and/or host proteins to remain soluble. To counter this insolubility, and to compete with host sequestration mechanisms, many unicellular pathogens will secrete low molecular weight, high-affinity Fe³⁺ chelators known as siderophores. Once acquired, unicellular pathogens must liberate the siderophore-bound Fe³⁺ in order to assimilate this nutrient into metabolic pathways. While these organisms may hydrolyze the siderophore backbone to release the chelated Fe³⁺, this approach is energetically costly. Instead, iron may be liberated from the Fe³⁺ - siderophore complex through reduction to Fe²⁺, which produces a lower-affinity form of iron that is highly soluble. This reduction is performed by a class of enzymes known as ferric reductases. Ferric reductases are broadly-distributed electron-transport proteins that are expressed by numerous infectious organisms and are connected to the virulence of unicellular pathogens. Despite this importance, ferric reductases remain poorly understood. This review provides an overview of our current understanding of unicellular ferric reductases (both soluble and membrane-bound), with an emphasis on the important but underappreciated connection between ferric-reductase mediated Fe³⁺ reduction and the transport of Fe²⁺ via ferrous iron transporters.

Graphical Abstract

Iron acquisition is crucial for the survival of unicellular organisms, particularly during pathogenesis. Many infectious organisms use ferric reductase enzymes to liberate Fe²⁺ from Fe³⁺ complexes, including siderophores and host proteins. The use of ferric reductases thus links ferric and ferrous iron transport.

Introduction

Since early in the history of life, unicellular organisms have incorporated iron-based cofactors into many proteins that execute essential cellular functions. These cofactors expand the repertoire of amino-acid based chemical reactions and add redox functionalities that are generally inaccessible to a protein-only scaffold [1]. Emphasizing the rich chemistry achieved by this essential element coordinated by a polypeptide, iron-containing proteins are known to synthesize DNA [2], to fix N₂ [3], to transport O₂ [4, 5], and to participate in photosynthesis [6], among other processes [1]. Thus, the acquisition and the management of iron are both necessary for the survival of most organisms. Unfortunately, the introduction of oxygen to the Earth’s atmosphere during the Great Oxygenation Event (GOE) drastically altered the availability of iron in these organisms’ environments [1, 7]. This crucial element, once readily abundant and soluble in the oceans, became sequestered within insoluble sediments [1, 7]. The newfound scarcity of accessible iron posed new challenges for the survival of most life forms.

Virtually every organism must acquire iron from its surroundings in order to maintain life. In the natural environment, iron typically exists in one of two oxidation states: 2+ (d⁶) or 3+ (d⁵). Under anoxic and/or acidic conditions, reduced ferrous iron (Fe²⁺) dominates [8–10]. Due to its modest molar solubility in water (up to ca. 10⁻³ M), ferrous iron may be readily acquired by organisms via strongly conserved membrane protein complexes [8, 9]. However, excess intracellular reduced iron can be exceedingly dangerous, as misregulated levels of such a reactive transition metal make organisms vulnerable to dangerous Fenton-like reactions that produce spuriously reactive hydroxyl radicals [1, 8, 11]. In contrast, under oxic conditions, oxidized ferric iron (Fe³⁺) ion is more prevalent [8–10, 12] but readily forms insoluble oxides in most aqueous environments (molar solubility of 10⁻¹⁸ M at pH 7), making acquisition significantly more difficult for organisms [8, 12]. If intracellular iron levels are too low, or iron mobilization from the environment too difficult, organisms cannot synthesize the key metalloproteins that are necessary for general cellular function, such as DNA biosynthesis, electron transport, and even aerobic cellular respiration [1, 8, 12]. Thus, organisms have evolved many means of acquiring this necessary element, especially at the host-pathogen interface (Fig. 1).

Figure 1.

Common Fe³⁺ acquisition strategies used by pathogens. While multiple Fe³⁺ acquisition strategies exist, three systems are prevalent and are pictured here for a Gram-negative bacterium. Left: the uptake of iron protoporphyrin IX (heme b. or “heme”) most commonly occurs via coupling of a TonB-dependent outer-membrane transporter to ATP-dependent ATP-binding cassette (ABC) transporters in the inner membrane. Middle: the uptake of Fe³⁺ siderophores (Fe³⁺-Sid) commonly occurs via coupling of a TonB-dependent outer-membrane transporter to ATP-dependent ABC transporters in the inner membrane. In cases where iron is released from the siderophore in the periplasm, different inner membrane transporters are used. Right: the uptake of non-chelated Fe³⁺ most commonly occurs via coupling of a TonB-dependent outer-membrane reception that usurps Fe³⁺ from iron-bound transferrin, which is translocated into the periplasm and chelated by the iron-utilizing periplasmic solute-binding protein FbpA that hands off Fe³⁺ to ATP-dependent ABC transporters in the inner membrane. Similar strategies are used for other pathogens that lack multiple membranes, commonly without the involvement of the outer-membrane transporters.

The procurement of iron is known to be a key virulence factor of unicellular pathogenic bacteria and yeast [1, 8], which have evolved multiple methods for acquiring this metal, particularly when faced with competition of the host organism (Fig. 1). During the early stages of infection in mammals, the mammalian host responds rapidly by sequestering its iron supply into iron-containing proteins (such as hemoproteins), chaperones (such as transferrin and lactoferrin), or larger storage proteins (such as ferritin) [13, 14]. This action results in a free serum iron concentration within the host of ca. 10⁻²⁴ M [14]. Nearly as rapidly, pathogens deploy additional measures to counteract their host’s iron sequestration methods. For example, some pathogens secrete proteins known as hemophores, which bind and import free heme and/or heme extracted from host hemoproteins (Fig. 1, left) [15, 16]. In pathogens with multiple membranes, active transport of heme across the outer membrane is typically driven by proton motive force (PMF) created by the inner membrane protein complex TonB/ExbB/ExbD (Fig. 1) that is present in multiple copies per cell [1, 16]. An ATP-binding cassette (ABC) transporter protein will typically then transport the heme across the inner membrane of bacteria (Fig. 1) [1]. In another example, iron-bound transferrin may be imported directly by some pathogenic microorganisms to combat iron sequestration directly (Fig. 1, right) [17, 18]. Proteins in the periplasm are capable of removing iron from the imported transferrin and delivering the iron to inner membrane transporter proteins (Fig. 1, right) [17]. Despite these common strategies, some microorganisms lack these capabilities and must rely on other tactics to acquire iron from these chaperones [19].

To compete with the chelating proteins found in host environments, and to address the challenge of the insolubility of ferric iron, many microorganisms will synthesize low molecular weight, high-affinity ferric iron chelating molecules known as siderophores (Fig. 2) [12, 18–21]. These multidentate chelators form soluble, highly stable Fe³⁺ complexes that can then be imported into the bacterial cell through pathways similar to those used for hemophores (Fig. 1, middle) [12, 18–21]. Siderophores are commonly oxygen-rich, Lewis-basic molecules that are loosely categorized based on the functional groups they use to chelate iron, such as hydroxamates, phenolates, catecholates, and even (α-hydroxyl-)carboxylates (Fig. 2) [12]. A given siderophore may use one or a mixture of these functional groups (Fig. 2) [12]. In many cases, microorganisms will maintain several membrane-imbedded transporter proteins in order to import multiple types of siderophores [9, 12, 18–21] (Fig. 1), and siderophore uptake systems have been identified as a possible target for antibiotic treatments [22, 23]. Furthermore, microorganisms are even capable of importing and using types of siderophores produced by other organisms, commonly known as xenosiderophores [12, 18–21]. This capacity is seen even in organisms that are not known to produce any siderophores, giving certain unicellular organisms a competitive edge in the battle over iron at the host-pathogen interface [18, 20, 21, 24]. However, as most Fe³⁺ siderophore complexes are both exceptionally thermodynamically and kinetically stable, an organism is then challenged with the task of releasing the iron from the chelate in order to assimilate it into the metabolic iron pool.

Figure 2.

The composition of common bacterial siderophores. The most common siderophore-based Fe³⁺-binding moieties (top row, left to right) are: catecholates, phenolates, hydroxamates, and (α-hydroxyl-)carboxylates. Example siderophores with these moieties (bottom row, left to right) are vibriobactin, pyochelin, desferrioxamine B, and staphyloferrin A. The example functional groups are highlighted yellow (catecholate), orange (phenolate), green (hydroxamate), and blue ((α-hydroxyl-)carboxylates) in each siderophore, respectively.

Once an Fe³⁺ siderophore complex is taken into the cell, organisms employ one of two common strategies for metal release. In one method, the siderophore is hydrolyzed to release Fe³⁺ [12]. While ultimately effective, this brute-force approach requires the catabolism of molecules that utilize significant energy and metabolites to assemble, which would represent a wasteful strategy [12, 17]. A more parsimonious approach is the reduction of Fe³⁺ siderophores to Fe²⁺ siderophores by a class of enzymes known as ferric reductases. This strategy weakens the binding affinity of the complex and results in iron release [9, 10, 12]; if the release occurs outside of the cytoplasm, the ferrous iron may either be transported into the cytoplasm via ferrous iron transporters[9, 10, 20] or re-oxidized for transport by ferric iron transporters [9, 10]. Importantly, this approach links Fe³⁺ assimilation to Fe²⁺ transport even under oxic conditions, connecting ferric and ferrous iron acquisition systems, an underappreciated but key strategy used by many pathogens to supply intracellular iron. This review will provide an overview of ferric iron reductases found in unicellular organisms, with a particular focus on the enigmatic and poorly understood class of membrane-bound reductases and their connection to iron acquisition. As these ferric iron reductases are common to many unicellular pathogens, this pathway could represent a viable target for small molecule development to attenuate the virulence of infectious unicellular organisms.

Soluble Ferric Reductases in Prokaryotes

The majority of bacterial ferric reductases that have been studied are soluble proteins that are either specific to certain types of ferric iron complexes [22, 25–28] or, more commonly, nonspecific flavin (ferric) reductases that use reduced flavins subsequently to reduce ferric iron [29–33]. Flavin cofactors, prosthetic or free, are used in most cases [25, 26, 29–34], but notable siderophore-specific reductases have been identified that use other cofactors [22, 27, 28]. These two varieties of reductases serve as complementary soluble pathways for extracting iron from the many types of siderophore complexes that the bacterium may encounter. One pathway, using reduced flavins, is broadly applicable to many different Fe³⁺-siderophore, Fe³⁺-citrate, and other Fe³⁺ complexes but is nonspecific. The other pathway uses specific binding of the enzyme to ferric complexes, primarily specific Fe³⁺-siderophore complexes, to modulate the high thermodynamic stability of these Fe³⁺ complexes, facilitating reduction and Fe²⁺ release.

Flavin (Ferric) Reductases

The majority of soluble ferric reductases in bacteria are flavin reductases that use a reduced flavin (e.g., FMNH₂; Fig. 3) to reduce a variety of substrates including Fe³⁺ complexes [10, 29, 34]. This process is sequential, requiring a separate cofactor to generate the reduced flavin initially. As a first step, these flavin reductases will typically use obligate two-electron reducing agents such as NADH [29, 30] or NADPH [25, 29, 31, 32] to reduce FAD [25, 26, 30, 31, 33], FMN [29], or riboflavin [32, 33] cofactors (Fig. 3), which may be prosthetic [25, 26, 29, 31, 34] or free [29, 30, 32, 33]. This process primes the flavin reductase, which may then reduce any of several Fe³⁺ complexes, provided the redox potential is favorable: the flavin/dihydroflavin redox couple is ca. −200 mV [25], whereas Fe³⁺ transferrin may have a redox couple ca. −300 to −500 mV [35, 36], and some Fe³⁺ siderophores may have potentials as low as −700 mV, too stable to be reduced by reductant molecules alone [37]. In general, flavin reductases are nonspecific in their ferric reductase activity and are capable of non-specifically reducing Fe³⁺-siderophore complexes, Fe³⁺ citrates, synthetic iron complexes such as Fe³⁺-EDTA, and even oxidized ferritins [29, 31–33] (Fig. 4). In contrast, however, some flavin reductases bind prosthetic flavins, which increases their affinity for particular siderophores; these reductases are still capable of broadly reducing Fe³⁺ complexes other than their presumed target substrates but with lower catalytic efficiency (e.g., 1.3x10³ s⁻¹ M⁻¹ vs. 2.3x10⁴ s⁻¹ M⁻¹) [25].

Figure 3.

The most frequent coenzymes that function in concert with soluble ferric reductases. Left panel. The reduced form of nicotinamide adenine dinucleotide bound to ferric reductases may either have a hydroxyl group at the 2’ position (NADH) or a phosphate group at the 2’ position (NADPH). Right panel. The flavin cofactors bound to ferric reductases all have the main isoalloxazine ring but differ in their substitution at the ribitol backbone: riboflavin is defined as R=OH; FMN is defined as R=PO₄³; and FAD is defined as R=adenosine diphosphate.

Figure 4.

Soluble ferric reductases. Left panel. The majority of soluble ferric reductases are flavin (ferric) reductases, soluble proteins that use pools of NAD(P)H to generate reduced, bound flavins. Once reduced, these flavin-bound reductases drive the reduction of Fe³⁺ siderophores (Fe³⁺-Sid) with modest reduction potentials to liberate Fe²⁺ and the intact siderophore. Right panel. Some soluble ferric reductases use an [Fe-S] cluster instead of a flavin cofactor to reduce Fe³⁺ siderophores. Currently, these soluble [Fe-S] (ferric) reductases are known to have a preference for hydroxamate-based siderophores (Fe³⁺-Sid_hyd). The source of electrons that generates the reduced [Fe-S] cluster is currently unknown.

The best-studied example of a bacterial nonspecific flavin reductase is the NAD(P)H:flavin oxidoreductase (Fre) from Escherichia coli [30, 32, 33], although homologs in other bacteria likely exist. Fre uses cellular NAD(P)H to reduce free FAD or riboflavin, which in turn has been found to reduce Fe³⁺ compounds within cells [30, 32, 33]. However, Fre reduction of FAD has been found to accelerate dangerous Fenton-like reactions through production of excess Fe²⁺ in respiratory-deficient cells that contain excess levels of NADH [30]. This spurious side reaction must be controlled, and it is thought that nonspecific reduction is kept in check by Fre’s high K_M for NADH (ca. 0.3 mM) when reducing FAD, which exceeds the NADH concentration in cells undertaking normal respiration (ca. 0.05 mM) [30, 33]. Flavins found to be reduced by Fre are known to be involved in the reduction of the di-metal cluster in ribonucleotide reductase (RNR) [33, 38], and in the reduction of many Fe³⁺-siderophore complexes [32, 33]. Although not strictly a flavoprotein per se, Fre shares structural homology to the flavoprotein family ferredoxin-NADP⁺ reductase (Fpr), which is known to use a bound flavin cofactor to reduce ferredoxin as well as siderophore-bound Fe³⁺ [29, 31]. Fprs strongly prefer NADPH as an electron donor to reduce their prosthetic flavin, although some have been shown to reduce iron more quickly in the presence of free flavin [29, 31].

A ferric reductase with homology to a family of NAD(P)H:flavin oxidoreductases has also been identified in archaea, from the hyperthermophile Archaeoglobus fulgidus [39, 40]. Known as FeR, this archaeal reductase has been shown to reduce exogenous, synthetic Fe³⁺ complexes and Fe³⁺-citrate using NAD(P)H and a bound FAD or FMN cofactor [39]. Although evidence has been found for iron reduction pathways in other archaea [41, 42], FeR from A. fulgidus is the only archaeal reductase to date that has been isolated and characterized [39, 40, 42]. It is not yet known whether FeR is involved in iron acquisition, some other function involving iron reduction, or both [39], and FeR’s affinity for reducing Fe³⁺-siderophore complexes has yet to be evaluated [39]. More work is clearly needed to gain a better understanding of archaeal flavin (ferric) reductases.

Siderophore-Specific Ferric Reductases

Of the flavoproteins in bacteria that reduce iron, there are some that contain binding pockets for certain siderophore complexes, and these are commonly termed siderophore-interacting proteins (SIPs). One of the best examples is the YqjH protein from E. coli [25]. YqjH accepts electrons exclusively from NADPH and binds Fe³⁺-tris(catecholate) siderophore complexes (K_d ca. 0.3-1.2 μM) and also Fe³⁺ dicitrate (K_d ca. 12 μM) complexes, but not hydroxamate complexes [25]. Intact tris(catecholate) complexes generally have extremely negative reduction potentials (ca. −750 mV); however, binding of these complexes by YqjH shifts the reduction potential, facilitating metal reduction [25]. Furthermore, preemptive hydrolysis of tris(catecholate) accelerates metal reduction due to a reduction of the chelate effect [25], but these strategies seem to be catecholate-specific in E. coli. Similar binding (K_d ca. 30 μM) is found in the FscN protein from Thermobifida fusca [26]. FscN uses a bound FAD cofactor and NADH to reduce the mixed catecholate-hydroxamate siderophore fusachelin A [26]. SIPs have also been identified that bind and reduce hydroxamates, such as SfSIP from Shewanella frigidimarina [34]. SfSIP may use either ferredoxin or NAD(P)H as an electron donor for reduction, although use of NAD(P)H requires the presence of an Fe²⁺ chelating agent [34]. Multiple hydroxamate siderophores produced by Shewanella species have been shown to be reduced by SfSIP [34]. However, not all reductases are dependent on NAD- or flavin-based cofactors to reduce specific Fe³⁺ siderophores.

In addition to the flavoproteins that are known to have affinity for specific siderophores, a handful of reductases have been identified that use a [2Fe-2S] cluster and reduce a specific class of Fe³⁺-siderophore complexes (Fig. 4). This family of proteins is typically called the ferric siderophore reductase (FSR) family [27]. One example is the FhuF protein, which is found in E. coli K-12 [27, 28] but is also prevalent in other pathogens. The Fhu system is generally responsible for the transport of the hydroxamate siderophore ferrioxamine B into the cytoplasm [27, 28]. The Fe³⁺ siderophore complex is translocated across the outer membrane by the TonB-dependent transporter FhuE [43] and brought into the cytoplasm by the ABC transporter complex FhuCDB [44] (Fig. 1). The reductase FhuF reduces Fe³⁺-ferrioxamine B and other hydroxamate siderophores in the cytoplasm via a [2Fe-2S]^2+/+ cluster bound within a unique binding motif (CCX₁₀CXXC), subsequently facilitating Fe²⁺ release [27] (Fig. 4). This reduction process has been shown to proceed in vitro with sodium dithionite serving as the initial cluster-reducing agent, although the exact electron source that reduces the cluster in vivo is not currently known [27] (Fig. 4). The binding of Fe³⁺-ferrioxamine B by FhuF is speculated to change the redox potential of the complex, which is too negative in solution (ca. −450 mV [37]) to be reduced by flavin reductases [27, 28]. While this change in potential has not been confirmed or characterized, it is thought to make the Fe³⁺ ion in the complex reducible by the Fe-S cluster (midpoint potential ca. −310 mV) [28]. A mechanism by which a specific siderophore is bound and its redox potential altered to facilitate iron reduction may serve to complement the nonspecific, broadly applicable but weak reduction potential of many flavin reductases. A similar protein named FchR has been identified in the Gram-positive alkaliphile Bacillus halodurans [22]. While it shares little sequence identity with FhuF, FchR binds a [2Fe-2S] cluster with a similarly unique Cys motif (CCX₄CX₆CXXC), has a similar midpoint potential (ca. −350 mV), and also reduces hydroxamate siderophore complexes [22]. The in vivo electron donor to FchR is also not known, but ferredoxin was found to be an effective electron donor in vitro [22]. It was further found that FchR could be competitively inhibited by the redox inactive Ga-siderophore complexes, which attenuated cell growth when supplemented in addition to hydroxamate siderophore complexes as an iron source [22]. These results suggest that FchR uses a similar strategy of binding ferric siderophores to facilitate iron reduction and release, which is incapable of occurring in the redox-inactive Ga complexes. Thus, while the exact electron donor may vary, some unicellular ferric reductases use tight and specific binding of Fe³⁺ siderophore complexes to raise the reduction potential of the siderophore-bound Fe³⁺ ion to more favorable ranges, facilitating metal reduction and release. This strategy represents one of a number of those employed as virulence factors to aid in the establishment of pathogenesis at the host-pathogen interface, and some infectious eukaryotes have evolved similar strategies.

Soluble Ferric Reductases in Yeast

In addition to bacteria and archaea, some unicellular eukaryotes such as pathogenic yeast make use of ferric reductases during infection [18, 45, 46]. Unlike in prokaryotes, soluble ferric reductases are much rarer in fungi, and only a handful have been identified and explored [18, 45]. To date, the characterized soluble fungal ferric reductases are typically excreted by the cell to reduce iron sources in the extracellular environment. Glutathione (GSH)-dependent ferric reductase activity has been identified in proteins excreted by the fungal pathogens Histoplasma capsulatum (cause of histoplasmosis, primarily a lung disease affecting immunocompromised individuals), Paracoccidiodes brasiliensis (cause of paracoccidioidomycosis, a broad infection affecting multiple organs), Blastomyces dermatitidis (cause of blastomycosis, primarily a lung disease), and Sporotherix schenckii (cause of sporotrichosis, primarily a skin disease) [47]. Isolation of the responsible glycoprotein from H. capsulatum revealed a two-step process for iron reduction [48]. and it is possible the mechanism may be operative in other organisms. First, the secreted enzyme known as γ-glutamyltransferase (Ggt1) hydrolyzes the γ-glutamyl-cysteine bond in GSH to release the dipeptide cysteinylglycine [48]. This step is known to be inhibited by classic Ggt inhibitors, such as acivicin, 6-diazo-5-oxo-L-norleucine (DON), and serine/borate complex (SBC) [48]. Second, and unique among Ggt enzymes, the cysteinylglycine dipeptide then reduces Fe³⁺ in the extracellular environment [48]. Although the exact Fe³⁺ species in vivo is unclear—Fe³⁺-nitrilotriacetic acid was used as a substrate in vitro—the reduction step has been found to be inhibited by the redox-inert iron analogues Al³⁺ and Ga³⁺ [48]. This demonstrated activity reveals an additional pathway for labilizing iron in a cell’s environment, operating in tandem with fungal siderophore-based pathways and reduction at the cell surface facilitated by membrane-bound ferric reductases.

Membrane-Bound Ferric Reductases in Unicellular Iron Metabolism

In addition to soluble ferric reductases, both pathogenic unicellular eukaryotes and bacteria use membrane-bound enzymes for iron reduction. Some of these enzymes function as part of a siderophore transport complex [49]. but many function as electron transporters across membranes to deliver reducing equivalents into a vacuole, the periplasm, or the extracellular space[18, 21, 45, 46, 50]. The intracellular source of electrons is typically NAD(P)H and/or flavins, similar to many of the soluble reductases found in bacteria [18, 49, 51]. However, transmembrane reductases are set apart from their soluble and membrane-anchored counterparts as many of them use heme b as the conduit for electron transport (Fig. 5), and hemes are known to span wide redox potential ranges, which could better match the potential of the ferric siderophore species as discussed below.

Figure 5.

Membrane ferric reductases. The majority of characterized membrane ferric reductases (fuchsia) are thought to be di-heme b proteins and are believed to belong to the Cytb₅₆₁ family of proteins. These membrane ferric reductases are generally small (predicted 2-6 transmembrane α helices) and may oligomerize. In pathogenic yeast, membrane ferric reductases are often larger due to the presence of multiple soluble domains that bind FAD and/or NAD(P)H. The source of electrons is thought to flow from NAD(P)H, through electron transport chain proteins in the membrane (red), to the membrane ferric reductases. After reduction of the Fe³⁺ siderophore (Fe³⁺-Sid), liberated Fe²⁺ is believed to be imported via a ferrous iron transporter (such as the bacterial FeoB protein, or the yeast Fet3/Fet4/Ftr1 complex; purple). A high-affinity Fe²⁺ transport system and/or an Fe²⁺ chelator (intra- or extracellular) may be essential to shift the equilibrium and make the reduction of the Fe³⁺-Sid more thermodynamically favorable.

Membrane-Bound Ferric Reductases in Bacteria

Many species of bacteria are known to have ferric reductase activity from proteins embedded in the cytoplasmic membrane [9, 49, 52–55]. In several studied examples, these reductases transfer reducing equivalents from reductants in the cytosol across the lipid bilayer, either into the periplasm or the extracellular space (Fig. 5) [21, 54–56]. These enzymes commonly function as stand-alone proteins based on previous studies. However, a membrane-tethered ferric reductase domain has been discovered as part of an unusual class of hydroxamate-phenolate siderophores transporters. In this section, we discuss these two classes of membrane ferric reductases: stand-alone reductases and membrane-tethered reductases that function in concert with siderophore importers.

A novel ferric siderophore reduction pathway was found in the ATP-binding cassette (ABC) transporter IrtAB, which is present in several Mycobacterial species such as M. tuberculosis and M. thermoresistible [49, 51]. IrtAB is a heterodimeric transmembrane protein, consisting of subunits IrtA (ca. 93 kDa) and IrtB (ca. 61 kDa), which are responsible for the import of the siderophores mycobactin and carboxymycobactin [49, 51, 57]. While each subunit contains an ATPase domain and a permease domain, IrtA contains an additional N-terminal domain that cradles a bound FAD cofactor [49, 51]. This N-terminal domain is termed the siderophore-interacting domain (SID) and has sequence homology to siderophore-interacting proteins (SIPs) in other bacteria [49, 51]. The SID has been shown to reduce both mycobactin and carboxymycobactin using at least NADPH as an electron donor after import by the transmembrane domain [49, 51]. This membrane-anchored SID is optimally positioned to reduce imported mycobactin, which is highly hydrophobic and remains tethered to the cytoplasmic membrane throughout its use [49]. The combination of active transport and reduction in one protein complex allows for rapid and efficient acquisition of iron.

In contrast to the Mycobacterial SIDs, most membrane-bound reductases in bacteria are stand-alone transmembrane proteins that shuttle electrons from cytosolic electron donors into the periplasm (Gram-negative bacteria) [21, 54, 55] or extracellular space (Gram-positive bacteria) [24, 58, 59]. Several studied reductases belong to a family of di-heme proteins known as cytochrome b₅₆₁ (Cytb₅₆₁) (Fig. 5). The Cytb₅₆₁ family is somewhat enigmatic but has a few known functions, including Fe³⁺ reduction [21] and superoxide oxidation [60]. These proteins are part of a larger superfamily of prokaryotic and eukaryotic proteins responsible for various electron transport functions across lipid bilayers [21, 45, 60–62]. Generally, Cytb₅₆₁ proteins contain two heme b molecules that act as electron conduits with the heme irons ligated each by two conserved His residues [21, 60] (Fig. 5). Surprisingly, the high potential heme (closest to the cytosol) has a redox couple ca. +150-200 mV, and electrons are subsequently transferred against the thermodynamic gradient to a low potential heme (near the substrate-binding site) that has a redox couple ca. 60 mV [61, 62] This “uphill” electron transfer is sometimes found in electron-transfer conduits, provided the overall reduction process is energetically favorable [63]. However, a major challenge remains to find the additional driving force necessary to reduce the highly stable Fe³⁺-siderophore complexes, which is currently unknown. Despite this apparent conundrum, ferric reductases still manage to reduce and to release iron in vivo.

A recently characterized example is the ferric reductase cytochrome b (FrcB) enzyme from Bradyrhizobium japonicum, an α-proteobacterium that lives in soil and is often found fixing nitrogen in symbiosis with common commercial crops such as soybeans [21]. B. japonicum is noteworthy in that it does not produce its own siderophores but is able to use xenosiderophores produced by fungi and other Pseudomonads found in its environment [21]. The frcB gene encodes for a transmembrane α-helical protein of ca. 20 kDa; further, frcB is found within a gene cluster containing a pyoverdine siderophore transporter gene, pointing to a role in iron uptake via xenosiderophore acquisition [21]. Electronic absorption spectroscopy confirmed the presence of two heme b cofactors, and both are necessary for the folding and stability of FrcB based on studies of variant proteins that fail to accumulate in the absence of heme or its Fe-binding ligands. FrcB has been shown to reduce FeCl₃ in vitro following reduction by dithionite [21], but the in vivo electron donor for FrcB is not yet known [21]. Given the connection of FrcB to iron acquisition and root nodule symbiosis, it is tempting to speculate ferric reductase activity is imperative for iron assimilation into nitrogenase and/or associated electron transport proteins.

Homologous Cytb₅₆₁ proteins are wider spread than simply symbiotic α-proteobacteria, with these putative reductases present in a system of proteins responsible for the use of pyoverdine siderophores in Pseudomonads, including the opportunistic P. aeruginosa [54, 56]. In P. aeruginosa the ferric reductase FpvG works in conjunction with 9 other proteins encoded by 3 different operons, in addition to a separate outer membrane efflux pump, to acquire iron from siderophores [54, 56]. FpvG is a ca. 46 kDa protein predicted to contain a total of 4 transmembrane α helices, a topology that strongly resembles FrcB [54]. Despite its sequence homology to Cytb₅₆₁ proteins, FpvG is not currently known to have a heme binding site [54], although the native reductive cofactor remains unknown. Fe³⁺-pyoverdine is believed to be shuttled to FpvG from outer membrane pores FpvA and FpvB by a complex of periplasmic binding proteins FpvF, FpvC, and FpvJ [64]. FpvG reduces the Fe³⁺-pyoverdine in complex with inner membrane protein FpvH and possibly inner membrane protein FpvK [54, 56]. The exact function of these other inner membrane proteins is not yet known, and the mechanism of this process remains unclear, but loss of the inner membrane proteins lowers FpvG activity [54, 56]. After reduction, dissociation of Fe²⁺ from pyoverdine is assisted by FpvC, which binds Fe²⁺ [65] and presumably delivers it to the ABC transporter FpvDE [54, 56], although whether P. aeruginosa’s Feo systems also contributes is unknown. FpvF then shuttles the empty pyoverdine to the PvdT protein involved in efflux into the extracellular space, completing the cycle of iron uptake in which FpvG takes part [66].

Another family of putative ferric reductases is the COG3295 family, which is also found in several Gram-negative pathogens. Of these, the VciB enzyme from Vibrio cholerae has been functionally characterized, as well as its homologs from Burkholderia mallei, Burkholderia thailandensis, and Aeromonas hydrophila [55]. A putative outer membrane transporter VciA is found encoded upstream from VciB in V. cholerae and A. hydrophila, but VciA is not believed to be necessary for VciB function [55]. Studies at the protein level have suggested VciB is a homodimeric enzyme, with each ca. 27 kDa monomer containing three transmembrane domains, a periplasmic loop, and two His residues essential for enzymatic function [55]. The total of 4 essential His residues per homodimer is consistent with the possibility that these proteins could bind heme similar to the Cytb₅₆₁ ferric reductase counterparts, but the presence of heme in VciB has not been elucidated [55]. Reduction of FeCl₃ by VciB has been demonstrated in vivo and is known to stimulate iron transport via the Feo system [55]. Gene deletion experiments suggest that the reducing equivalents used by VciB come from upstream NADH-dependent proteins in the membrane electron transport chain [55]. This connection, along with similar evidence from other cytochrome b-type enzymes [60], provides a promising lead for understanding other Gram-negative membrane-bound reductases for which the in vivo electron donor is not known.

Less is known about the role of membrane-bound reductases and their connection to iron metabolism in Gram-positive bacteria. Ferric reductase activity has been localized to the cell surface of Listeria monocytogenes and is linked to the virulence of this Gram-positive pathogen [24, 59], but it is not yet clear whether a membrane-bound or secreted enzyme is responsible [24]. Moreover, the composition of these ferric reductases is not well studied, with cytochrome c-type enzymes and flavin-binding lipoproteins having been implicated in the transfer of electrons to iron and other metal substrates at the cell surface; however, this activity is thought to have a purpose outside of metal uptake [58, 63, 67–69], and it is unclear whether this electron transport stems from a bona fide ferric reductase. This process, commonly referred to as extracellular electron transfer, is capable of reducing iron compounds such as Fe³⁺-EDTA and Fe³⁺-oxides but is thought to use these compounds for anaerobic respiration rather than as an iron source [58, 63, 67–70]. More research is needed to determine which membrane-bound reductases, if any, may contribute to iron metabolism in Gram-positive bacteria, and exactly how this mechanism may differ from, or be similar to, the better-studied Gram-negative systems.

Membrane-Bound Ferric Reductases in Yeast

Yeast make extensive use of membrane-bound reductases, both in the plasma and vacuolar membranes. Evidence has shown that fungal membrane-bound ferric reductases make use of NAD(P)H, flavin, and heme b cofactors to deliver reducing equivalents across the lipid bilayer to a variety of Fe³⁺ species, depending on the reductase in question [18, 45, 46]. Emphasizing the importance of this metal-release pathway, yeast genomes will encode anywhere from 2 [71] to 17 [72] putative ferric reductases to reduce various Fe³⁺ substrates under different conditions. After electron transport, the newly reduced Fe²⁺ that is liberated from any number of chelators may be transported by low-affinity proteins that can transport Fe²⁺ (e.g., Fet4, Ccc1) [45]. However, a more common approach that is distinct in pathogenic eukaryotes is the reoxidation of the liberated iron, which is bound and transported across the membrane by a complex of the Cu-dependent ferroxidase (e.g., Fet3) and Fe³⁺ transport (e.g., Ftr1) proteins, [45, 73].

One of the most thoroughly characterized fungal reductases is the Fre family (commonly numbered Fre1 to Fre10), which have been studied extensively from the non-pathogenic Saccharomyces cerevisiae, but are present also in many problematic fungal pathogens such as Candida albicans [50, 74], Cryptococcus neoformans [18], and Aspergillus fumigatus [71]. Fres vary in size, from as small as 426 amino acids (ca. 48 kDa) to as large as 1032 amino acids (ca. 114 kDa), and as many as 9 transmembrane α helices are predicted with several soluble domains interspersed amongst the α helices; these domains (membrane and soluble) have sequence homology to heme-, FAD, and NAD(P)H-binding proteins [18]. In S. cerevisiae, over 90% of cell surface ferric reductase activity comes from the membrane-bound Fre1 and Fre2, which reduce a wide variety of Fe³⁺ substrates, including Fe³⁺-citrate as well as catecholate and hydroxamate siderophores (Fig. 2) [46]. Likewise, an analog of Fre2 in A. fumigatus known as FreB [75] is the primary ferric reductase for reductive iron uptake in this pathogen [71]. Fre2 homologs in C. neoformans have been shown to be virulence factors and play an important role in acquiring iron from transferrin [18]. In addition, although its function does not affect heme uptake across the plasma membrane, C. neoformans Fre2 has been shown to promote cell growth when low concentrations of hemin are used as an iron source [18]. Fre3 appears to be optimized to reduce hydroxamate siderophores such as ferrioxamine B and ferrichrome (Fig. 2) and can utilize these siderophores sufficiently to maintain cell growth in mutant strains deficient in Fre1 and Fre2 [46]. Fre4 is thought to be involved in the reduction of Fe³⁺-rhodotorulic acid, but not other hydroxamate siderophores [46]. Fre6 is localized to vacuolar membranes and is involved in mobilizing and exporting iron from Fe³⁺-polyphosphate stores in the vacuole [73]. Unsurprisingly for a protein embedded in a smaller membrane, Fre6 is one of the smaller Fre enzymes (ca. 48 kDa in C. neoformans) and contains only 1 putative transmembrane α-helical domain [18]. Fre10 is responsible for a majority of the ferric reductase activity in C. albicans under neutral and acidic conditions, and for the acquisition of iron from host transferrin [50]. Intriguingly, some Fre enzymes are also well known to reduce Cu²⁺, likely to support Cu-dependent proteins; reflective of this cross-talk, these ferric reductases appear to be transcriptionally regulated by both Fe and Cu levels in the cell [18, 45, 46]. Furthermore, evidence from a Fre4 homolog in C. neoformans also suggests a role in Cu uptake, which is critical for maintaining the cell’s ferroxidases and melanin production [18]. These observations suggest that the Fre family of ferric reductases may be essential virulence factors that connect both Fe and Cu homeostasis in pathogenic fungi.

C. albicans contains a set of ferric reductases sometimes known as C. albicans ferric reductase-like (Cfl) proteins [76], which have also been classified as Fre proteins although these are likely distinct in function [74, 77]. Cfl 1 shares loose sequence homology with the Fre1 and Fre2 enzymes in S. cerevisiae, is similar in size (ca. 85 kDa), has predicted cofactor binding sites (FAD, NAD(P)H, heme), and was the first ferric reductase identified in this pathogen [76]. In a separate study in non-pathogenic yeast, iron transport in Fre1-deficient (Δfre1) strains of S. cerevisiae can be restored by the addition of the cfl1 gene [76]. Cfl1 has known implications for controlling intracellular iron supply, resisting oxidative stress, and maintaining cell wall integrity, and deletion of this gene has been shown drastically to upregulate the expression of other reductases in this pathogen [72, 78]. An Fre2 homolog in C. albicans is sometimes called Cfl2, and this reductase may not be as essential as Cfl1 [74]. Both this homolog and another reductase, Frp1, are upregulated under more alkaline conditions [74, 77]. Although little is known about Cfl2 at the biophysical and biochemical levels, Frp1 is known to be the primary ferric reductase in the fission yeast Schizosaccharomyces pombe [79]. Despite being very distant phylogenetically, S. cerevisae and S. pombe share sequence homology in their heme-, FAD-, and NAD(P)H-binding domains of Fre1 and Frp1, respectively [79]. Similar domains are even found in a component of human NADPH phagocyte oxidoreductase [79]. The conservation of these domains across evolutionary space suggests a common NAD(P)H-to-FAD-to-heme b mechanism of electron transfer across lipid bilayers in unicellular eukaryotes [79]. Intriguingly, this large variety of reductases that are expressed under slightly altered conditions and tuned to different Fe³⁺ substrates gives these organisms great flexibility in acquiring iron within many different environments. This task is especially important for pathogenic fungi that do not produce their own siderophores, making ferric reductases likely important in the establishment of infection within a host organism. Despite this link to pathogenesis, these reductases have only been chiefly characterized at the genetic and cellular levels, and more functional and structural characterization at the protein level is needed to shed light on these crucial pathways.

Connecting Ferric and Ferrous Iron Transport

For both soluble and membrane-bound ferric reductases, it has been shown that the chelation of Fe²⁺ by other chaperone molecules has a significant impact on the rate of reduction, and this observation implies that the involvement of high-affinity Fe²⁺ transporters and/or chelators may help drive the liberation of iron from siderophores even under oxic conditions (Fig. 5). This phenomenon occurs because the reduction potential of the Fe³⁺ siderophore complex (generally very thermodynamically unfavorable) is dependent on the equilibrium concentrations of both the Fe³⁺ complex and the liberated Fe²⁺ species. Removal of free Fe²⁺ from solution shifts the reduction potential to more positive (and therefore more thermodynamically favorable) values, in keeping with the Nernst equation:

$$E=E^{°}-\frac{\mathit{RT}}{\mathit{nF}}\ln \left(\frac{\left[{\mathit{Fe}}^{2+}\right]}{\left[{{\mathit{Fe}}^{3+}}_{\mathit{complex}}\right]}\right)$$

Several examples of this equilibrium/potential shift have been observed. For instance, synthetic ferrous iron chelators such as ferrozine are known to stimulate significant Fe³⁺ reduction in vitro above basal levels with reductant alone [32, 34, 54]. Additionally, although this compound has not had much study, a phosphorylated sugar compound called ferrochelatin is speculated to serve a similar function in vivo, driving reduction by chelating soluble ferrous iron from within the cell [27]. For membrane-bound reductases, Fe³⁺ reduction is thought to be coupled to proteins that transport Fe²⁺ into the cytosol [9, 55, 80]. In bacteria, the most common Fe²⁺ transporter is the Feo system [8, 81]. The Feo system consists of FeoB, a high-affinity Fe²⁺ membrane transporter protein with a cytosolic GTPase domain [81]. and two cytosolic proteins, FeoA [82] and FeoC [83]. Despite it being a dedicated Fe²⁺ transport system, Feo is present in many pathogens undertaking both aerobic as well as anaerobic metabolism. In fact, deletion of feo genes has shown dramatic impacts on the growth and virulence of organisms undertaking aerobic metabolism with native and xenosiderophores present as iron sources [84, 85], and the presence of an intact feo operon is known to be important for extracellular reduction [19], which has been linked to ferric reductase activity in some cases [55]. Additional complementary pathways in bacteria [9, 55, 80] and systems such as the Fet3/Fet4/Ftr1 set of proteins in yeast [45, 73] are known to be necessary for the acquisition of iron through transmembrane reductases. We speculate that these Fe²⁺ transport pathways are present under oxic conditions (aerobic metabolism) to serve as Fe²⁺ sinks that shift the equilibrium (and thus the redox potential) of these highly thermodynamically stable Fe³⁺ siderophore complexes to more favorable reduction potentials for iron liberation (Fig. 5). While more work is undoubtedly needed, these observations suggest a key link between the reduction of ferric iron and the mobilization of ferrous iron in the cell, which is an exciting avenue of future research.

Outlook

It is critical for unicellular organisms to acquire iron, especially during pathogenesis. These infectious organisms frequently scavenge for iron in their environment by using native siderophores or by pirating xenosiderophores in order to bind and to solubilize Fe³⁺. One major route of Fe³⁺ liberation from chelation is via reduction to Fe²⁺ by ferric reductase enzymes. Both soluble and transmembrane enzymes are used to transfer electrons from intracellular reductants to Fe³⁺ complexes in the cytoplasm, periplasm, and/or extracellular space. This reduction is driven in part by the mobilization of the produced Fe²⁺, either by intracellular chelators or by Fe²⁺ transport proteins, strongly suggesting an important connection between ferric and ferrous iron transport even under oxic conditions. Both bacteria and yeast encode many reductase enzymes in their genomes, highlighting the importance of these enzymes for unicellular iron metabolism. Historically, the soluble, FAD-dependent ferric reductases have garnered much attention; however, there is much to be explored about these enzymes, especially for those that are heme b-dependent and membrane-bound, which appear to support higher redox potentials than those that are soluble. In particular, only a handful of Gram-negative bacterial membrane-bound reductases have been characterized, and none have been confirmed to participate in Gram-positive bacterial iron uptake. Likewise, little structural information is known about the membrane-bound reductases common to pathogenic yeast outside of a few sequence-based predictions. Less still in known about the role of ferric reductases in archaea, with only one enzyme characterized to date. Thus, it is clear that ferric reductases have many opportunities for future research, and a greater understanding of these systems could lead to the development of therapeutics to target these important electron-transport proteins for new classes of antibiotics.

HIGHLIGHTS.

• Ferric reductases are enzymes that reduce Fe³⁺ complexes to Fe²⁺ complexes

• Ferric reductases contribute to the virulence of problematic unicellular pathogens

• Soluble ferric reductases are NAD- and FAD-dependent soluble proteins

• Membrane ferric reductases are unique heme-dependent proteins

• Ferric reductases link ferric and ferrous iron acquisition

Abbreviations

ABC transporter

ATP-binding cassette transporter

ATP

adenosine triphosphate

DNA

deoxyribonucleic acid

FAD

flavin adenine dinucleotide

FMN

flavin mononucleotide

GSH

reduced glutathione

GTP

guanosine triphosphate

NAD(P)H

reduced nicotinamide adenine dinucleotide (phosphate)

SID

siderophore-interacting domain

SIP

siderophore-interacting protein