Structures and coordination chemistry of transporters involved in manganese and iron homeostasis

Abstract

A repertoire of transporters plays a crucial role in maintaining homeostasis of biologically essential transition metals, manganese and iron, thus ensuring cell viability. Elucidating the structure and function of many of these transporters has provided substantial understanding into how these proteins help maintain the optimal cellular concentrations of these metals. In particular, recent high-resolution structures of several transporters bound to different metals enable an examination of how the coordination chemistry of metal ion-protein complexes can help us understand metal selectivity and specificity. In this review, we first provide a comprehensive list of both specific and broad-based transporters that contribute to cellular homeostasis of manganese (Mn²⁺) and iron (Fe²⁺ and Fe³⁺) in bacteria, plants, fungi, and animals. Further, we explore the metal-binding sites of the available high-resolution metal-bound transporter structures (Nramps, ABC transporters, P-type ATPase) and provide a detailed analysis of their coordination spheres (ligands, bond lengths, bond angles, and overall geometry and coordination number). Combining this information with the measured binding affinity of the transporters towards different metals sheds light into molecular basis of substrate selectivity and transport. Moreover, comparison of the transporters with some metal scavenging and storage proteins, which bind metal with high affinity, reveal how the coordination geometry and affinity trends reflect the biological role of individual proteins involved in homeostasis of these essential transition metals.

Introduction

All living cells require transition metals in low intracellular concentrations for various biological processes. Cellular metal levels are tightly regulated as imbalances in their optimal concentrations cause several diseases [1,2]. In this review, we focus on two abundant and essential transition metals, manganese and iron, neighbors on the periodic table and thus closely related. Manganese is a cofactor of vital enzymes including superoxide dismutase (SOD) which breaks down reactive oxygen species (ROS) and glutamate synthetase involved in mammalian brain development and function [3,4]. Manganese is also important in processes like bone and tissue growth, reproduction, and immune system function [5,6]. Manganese overload in the brain leads to a condition called ‘manganism’ with Parkinson-like symptoms [6-8]. In plants, manganese catalyzes photooxidation of water by photosystem II during light reactions and provides structural integrity to various photosynthetic proteins and enzymes [9-11]. Manganese is also crucial for bacterial survival and growth, and virulence of pathogenic bacteria [12,13]. During pathogenesis, manganese (with or without SOD) protects the microbes from host-mediated oxidative stress by detoxification of ROS [4,14].

Iron, in ionic form or as part of a complex, is an essential cofactor of proteins involved in many processes including oxygen transport and storage, electron transport and catalysis, cellular and oxygen metabolism, DNA synthesis and repair, cellular signaling, and host defense [15-17]. Iron deficiency causes restricted erythropoiesis and anemia [17-20]. Elevated iron is also problematic, for example in organs it causes tissue damage (hemochromatosis), in the blood it leads to atherosclerosis, and in the brain causes neurodegeneration [17,21,22]. In cyanobacteria, algae, and plants, iron is the predominant metal in the photosynthetic machinery, and it contributes to electron transport and chlorophyll biosynthesis [23-25]. In stress conditions, iron imbalances generate ROS and impair plant growth, photosynthesis, electron transport [10,25,26].

Organisms have evolved tightly regulated systems to maintain optimal metal concentrations in cellular compartments. The metal homeostasis machinery includes: (i) transporters and ion channels that import and export metals across otherwise impermeable cellular and organellar membranes, (ii) metallochaperones and chelators that store and retain metals in certain compartments, and (iii) transcription factors, sensors, and riboswitches that regulate the expression of genes encoding for the homeostasis machinery [27,28]. Metal homeostasis necessitates that this machinery can select appropriate metals over others [27-29]. The coordination chemistry, geometric preferences, and chemical properties—like atomic and ionic radii, electronegativity—of individual metals are key contributors to this specificity [2,27,30]. However, according to the Irving-Williams series (Zn²⁺ < Cu²⁺ > Ni²⁺ > Co²⁺ > Fe²⁺ > Mn²⁺), a transition metal ion higher in the series tends to form more kinetically or thermodynamically stable protein-metal complexes and can potentially displace metals that are downstream in the series [31,32]. Hence, transition metal transporters often have a broad substrate scope even when physiologically implicated in homeostasis of a particular metal [27,28,33,34].

In this review, we address the factors driving selective iron and manganese transport and homeostasis. We first explore the transporters and receptors that contribute to manganese and iron homeostasis in mammals, yeasts, plants, and bacteria. From the list, we analyze the structures of different manganese- and iron-bound proteins available at high resolution to understand how structures correlate to biological function and factors that drive metal ion selectivity. Detailed analysis of the metal coordination spheres and their deviation from ideal parameters for different bound metals illustrates how the protein discriminates its cognate substrates from non-physiological ones. Reviewing the published affinities of metals to different proteins (either K_m or K_d), we find that they mostly, but not always, correlate to the ideality of the coordination geometries. Overall, we show that the combined knowledge of coordination structures and binding affinity can help understand the function of the key players involved in manganese and iron homeostasis.

Overview of manganese transporters in different organisms

Before we discuss the structural analyses, we briefly introduce the families of receptors and transporters—importers that bring substrates into the cytosol and exporters that ferry substrates from the cytosol to lumenal compartments or the extracellular space—implicated in manganese homeostasis. These players in manganese homeostasis are illustrated in Figure 1, listed in Table 1, and described by organism type below.

Figure 1. Transporters and receptors involved in manganese transport and homeostasis from bacteria to mammals.

(A) A gram-negative bacterial cell surface is shown as a representative for gram-positive and gram-negative bacteria because most Mn²⁺ transporters are similar in both. Mn²⁺ is imported across the OM of gram-negative bacteria mainly through porin channels like MnoP [35,161]. The illustrated ABC transporters are a representative, but not exhaustive, list (see Table 1). (B) A diagram of a plant cell highlighting transporters on the cell surface and membranes of internal organelles that regulate intracellular Mn²⁺ levels. YSL (Yellow Stripe-Like) proteins transport Mn²⁺-chelates, all other transporters transport Mn²⁺ ions [9,11,44]. (C) A diagram of a fungal cell highlighting transporters in the plasma membrane and organellar membranes which impact Mn²⁺ homeostasis. Mn²⁺ ions internalized into mitochondria primarily bind MnSOD (manganese-dependent superoxide dismutase), which protects against oxidative stress [2,4,52,53]. Most transporters shown have been more extensively studied in yeasts [2,51,52]. (D) In a typical mammalian cell, many plasma-membrane transporters control the cellular Mn²⁺ levels. The transferrin receptors (TfR) capture Mn³⁺-bound transferrin and internalize it into endosomes where Mn³⁺ is converted to Mn²⁺ and transported into the cytosol by NRAMP2 [66,70,71,164]. Several transporters—some of which are not solely or even primarily dedicated to Mn²⁺ transport—have been implicated in Mn²⁺ homeostasis in the brain, including NRAMP2, ZIP8/14, ZNT10, SOC, TfR, NMDAR, glutamate transporters GLAST and GLT-1, HIP14/14L, ATP13A2, DAT, MCT [70,160,164]. In each panel, specific proteins of each transporter family are shown, and their corresponding family is indicated in the legend at the bottom (see Table 1).

Table 1.

Transporters involved in manganese homeostasis

Family	Organism	Transporter	Function	Representative structurea
Nramp	Bacteria	MntH [12,13,34,136]	Mn2+ import	8E60
	Plants	NRAMPXb [10,47,137,138]	Mn2+ import at PM and vacuoles of roots and leaves	8E60
	Fungic	Smf1/2 [34,55,136]	Mn2+ import across PM and ER membrane	8E60
	Animals	NRAMP2 [9,34,66,136,139]	Mn2+ import into cytosol through PM and endosomes	8E60
	Animals	NRAMP1 [34,136]	Import Mn2+ into cytosol at phagosomal membrane	8E60
ABC transporterd	Gram-negative bacteria	SitABCD [97,140], MntABC [89]	Mn2+ import into cytosol and full virulence	4OXR
	Gram-positive bacteria	MtsABC [141], SloABC [142], PsaABC [31,74,143], EfaCBA [144], MntABC [13,33,145,146], SitABC [75]	Mn2+ import into cytosol and virulence	3ZTT
CDF	Bacteria	MntE [147], MneP/S [13,148], EmfA [103], YiiP [36,37]	Mn2+ export	2QFIe
	Plants	MTP1 [9,47,109]	Mn2+ export from cytosol into vacuole of roots, leaves	6XPEe
	Animals	ZNT10 [67,92,139]	Mn2+ export across PM in liver, intestine, and brain	6XPEe
P-type ATPase	Bacteria	MgtA, YoaB [12,13]	Mn2+ export	7YAJ
	Plants	ECA1, ECA3 [9,47,109,149]	Mn2+ uptake in ER (ECA1) and Golgi (ECA3) of roots	7YAJ
	Fungic	Pmr1p [2,51,52,55], Ypk9p [51,59], Cod1p [51,150]	Mn2+ export from cytosol into Golgi (Pmr1p), vacuole (Ypk9p) and ER (Cod1p)	7YAJ
	Animals	SPCA1/2 [8,69]	Mn2+ export from cytosol into Golgi	7YAJ
	Animals	ATP13A2 [3,68,71,139,151]	Mn2+ export into lysosome in brain	7VPIf
ZIP	Plants	IRT1 [47,109], ZIP1/2 [11,45]	Mn2+ import into cytosol from PM and vacuole (ZIP1) in roots	5TSBg
	Fungic	Atx2 [51,57,58,152]	Mn2+ import into cytosol from Golgi	5TSBg
	Animals	ZIP8, ZIP14 [3,67,71,139]	Mn2+ import across PM in liver and brain	5TSBg
CCC1	Fungic	Ccc1 [9,55,153]	Mn2+ export from cytosol into vacuole	6IU5e
	Plants, Fungic	VIT1/2 [48-50,138]	Mn2+ export from cytosol into vacuoles and Golgi	6IU5e
MFS transporter	Bacteria, Animals	ferroportin [71,77,116,154]	Mn2+ export	5AYMh
	Fungic	Pho84 [51,53,56,155]	Mn2+ import into cytosol across PM	7SP5f
	Animals	MCT [71,156-158]	Mn2+-citrate complex import into cytosol in brain	7JSJf
Cation exchanger	Plants	CAX2/4/5 [46,47,159]	Mn2+ export out of cytosol into vacuole	4KPPi
	Fungic	Gdt1p [51,60,61]	Mn2+ export out of cytosol into Golgi	4KPPi
	Animals	TMEM165 (36, 66, 67)	Mn2+ export out of cytosol into Golgi	4KPPi
Ca2+ channel	Animals	SOC [3,66,71,160]	Mn2+ import across PM in the brain	3JBRi
Porin	Gram-negative bacteria	MnoP [35,161]	Mn2+ import through OM	2PORi
MntP type	Bacteria	MntP [12,13,43,148]	Mn2+ export	-
TerC type	Bacteria	YkoY [12,13,38]	Mn2+ export	-
Oligopeptide transporterd	Plants	YSL [9,11,44,53,162]	Complexed manganese import at chloroplast, vacuole, and PM	-
CYSTM	Fungic	MNC1 [55,163]	Mn2+ chelation and import into cytosol	-
Transferrin receptor	Animals	TfR [3,66,70,71,164]	Internalization of Mn3+-transferrin	3S9Lh
Glutamate receptor	Animals	NMDAR [165-167]	Mn2+ import into cytosol across blood-brain barrier	5UP2f
EAAT	Animals	GLAST [165,168], GLT-1 [165,169]	Mn2+ import that causes neurotoxicity	6X12f
Palmitoyl acyltransferase	Animals	HIP14/14L [3,170,171]	Mn2+ export from cytosol into Golgi in the brain	3EU9f
NSS	Animals	DAT [71,172,173]	Mn2+ import into cytosol across PM in the brain	4XP1f
Mitochondrial carrier family transporter	Fungic	Mtm1 [52,62]	Mn2+ import from cytosol into mitochondria	1OKCe
	Mammals	MFRN1/2 [28,72]	Mn2+ import from cytosol into mitochondria	1OKCe

^aThe most relevant representative structure, prioritizing high-resolution structures with bound Mn²⁺, then ones with other metal ions, then ones with no bound metal.

^bX is a number; plants generally have 5-7 numbered NRAMP homologs.

^cFungi transporters are best characterized in yeasts, so yeast proteins are generally listed in this table.

^dNon-exhaustive list.

^eStructures with bound Zn²⁺.

^fStructures with no bound divalent ions at sites for transition metal transport.

^gStructure with bound Cd²⁺.

^hStructure with bound iron.

ⁱStructure with bound Ca²⁺.

Bacteria

Most importers and exporters contributing to manganese homeostasis are ubiquitously present in both gram-positive and gram-negative bacteria (Figure 1A, Table 1). Mn²⁺ importers include ATP-binding cassette (ABC) transporters, such as SitABC and PsaABC, that use energy derived from ATP hydrolysis, and a Natural resistance associated macrophage protein (Nramp) homolog, MntH, that uses co-transported protons as an energy source [12,33,34]. In Bradyrhizobium japonicum, a porin, MnoP, is coregulated with MntH and assists in Mn²⁺ diffusion across the outer membrane (OM) during Mn²⁺ deprivation [27,35]. Mn²⁺ exporters include: cation diffusion facilitator (CDF) family H⁺ antiporters like MntE, MntP, and YiiP [36,37]; P_II -type ATPases like YoaB and MgtA [12,13,27]; TerC-type proteins like YkoY [12,13,38]; MntP-type proteins; and the major facilitator superfamily (MFS) transporter FPN, a homolog of human ferroportin [39-41]. To maintain homeostasis, the expression levels of transporter genes are regulated by either or both riboswitches like yybP-ykoY [12,42] and transcription factors like MntR [12,43].

Plants

In plants, many transporters contribute to Mn²⁺ uptake through the roots and distribution in cells of various tissues and intracellular organelles (Figure 1B, Table 1) [9,11]. Mn²⁺ importers include Nramps like Oryza sativa (Os)Nramp3/5, Arabidopsis thaliana (At)Nramp1 (plasma membrane; PM) and AtNramp3/4 (vacuole) [9-11]. Yellow stripe-like (YSL) importers, belonging to oligopeptide transporter (OPT) family, mediate long-distance uptake of Mn²⁺ in complex with nicotianamine (e.g., AtYSL4/6, OsYSL2/6) [9,44]. Zrt-/Irt-like proteins (ZIPs) have a broad substrate scope including Zn²⁺, Fe²⁺, and Mn²⁺, with IRT1 (PM), ZIP1 (vacuole), and ZIP2 (PM) most strongly associated with Mn²⁺ import in A. thaliana [9,11,45]. The exporters that internalize Mn²⁺ in intracellular compartments include: cation exchangers (CAX) expressed in vacuoles in plants exposed to toxic Mn²⁺ levels (e.g., AtCAX2/4/5) [9,46]; CDFs (also known as MTP in plants; e.g., AtMTP11, OsMTP8.1) [9,47]; P-type Ca²⁺-ATPases like AtECA1 (endoplasmic reticulum; ER) and AtECA3 (Golgi) [9-11]; and vacuolar iron transporter (VIT) subfamily of Ca²⁺-sensitive cross complementer1 (CCC1) transporters in vacuoles and the Golgi (e.g., ATVIT1, OsVIT1/2) [48-50].

Fungi

The fungal Mn²⁺ homeostasis machinery has been most extensively studied in yeasts (Figure 1C, Table 1) [2,51]. Mn²⁺ importers include: Nramps like Smf1 (PM) and Smf2 (intracellular vesicles and ER) [52-54]; Mnc1, a CYSTM (non-secreted cysteine-rich peptide) family member that chelates manganese at the cell surface [55]; Pho84, a phosphate/proton MFS symporter which transports Mn²⁺ as MnHPO₄ in high environmental Mn²⁺ conditions [56]; and a ZIP transporter, Atx2, traffics Mn²⁺ out of Golgi vesicles [51,57,58]. To limit cytosolic Mn²⁺, several exporters internalize Mn²⁺ into intracellular compartments: P-type ATPases Pmr1 (Golgi), Ypk9 (vacuole) and Cod1 (ER) [52,54,59]; CCC1 exporters Ccc1 (vacuole) [52-54]; CAX exporter Gdt1 (Golgi) [60,61] and the mitochondrial carrier (MC)-family exporter Mtm1 (mitochondria) [62].

Animals

Animals use manganese (Mn²⁺ or Mn³⁺) homeostasis machinery for several essential life processes and for nutritional immunity against invading pathogens (Figure 1D, Table 1) [7,28,39]. The immunity proteins include the Nramp-family namesake, NRAMP1, which extrudes Mn²⁺ and Fe²⁺ out of the phagosomes of macrophages into the cytosol to deplete the engulfed pathogens of essential nutrients [34,39,63]. Mn³⁺ is carried in the bloodstream by transferrin (whose preferred substrate is Fe³⁺) then internalized into endosomes by the transferrin receptor (TfR), where Mn³⁺ is released and converted to Mn²⁺ by STEAP metalloreductases [34,64,65]. The main Mn²⁺ importers are NRAMP2, involved in uptake from endosomes and the PM [66], and ZIP8 and ZIP14, importing Mn²⁺ preferentially over Zn²⁺ and Fe²⁺ across the PM [28,67]. The main Mn²⁺ exporters are: P-type ATPases ATP13A2 (Park9; lysosomes) [28,68] and SPCA1 (Golgi) [28,69]; CDF-family transporter ZNT10 (Zinc transporter 10) that preferentially exports Mn²⁺ over Zn²⁺ across the PM [28,53,67]; CAX-family TMEM165 (Golgi) [28,60,70]; and the MFS transporter, ferroportin [28,71] and the MC-family MFRN1 and MFRN2 that import Mn²⁺ into the mitochondria in Fe²⁺ deficiency conditions [28,72].

Manganese is vital to brain function and its dyshomeostasis induces neurological disorders [6,28,39]. Accordingly, the generic cell in Figure 1D also includes many channels and transporters that primarily shuttle other substrates but have been implicated in Mn²⁺ fluxes in brain tissues (reviewed in [6,73]). These include store-operated Ca²⁺ channels (SOCs), voltage-gated calcium channels, ionotropic glutamate receptors (NMDARs), the electrogenic amino acid transporter (EAAT) family of glutamate and glutamine transporters, the dopamine transporter (DAT), MFS-family monocarboxylate transporter (MCT), and Huntingtin-interacting proteins HIP14 and HIP14L (Mg²⁺ transporters expressed in the Golgi).

Structural insights into manganese transport

Transporters involved in Mn²⁺ homeostasis either use Mn²⁺ as their primary substrate but can also transport other transition metals (e.g., MntH, PsaABC, SitABC) [74-76], or moonlight as Mn²⁺ transporters, as do most Fe²⁺ transporters (e.g., TfR, ferroportin), sometimes under dyshomeostasis conditions [41,64,77], and some Zn²⁺ transporters (ZNT10, ZIP8/14) [67]. In both cases, it will be useful to understand what drives Mn²⁺ selection in a physiological context. Mn²⁺ prefers an octahedral geometry, with optimal bond lengths of 2.1-2.5 Å as ‘strong interactions’, although longer bond lengths of 2.5-3.0 Å can still participate in a coordination sphere as ‘weak interactions’ [78-80]. However, chelation of Mn²⁺ can be constrained by the protein structure to differ from ideal coordination geometry; it is therefore important to visualize Mn²⁺-interactions to understand how it influences function. Although several Mn²⁺ transporter structures are available, few are high-resolution metal-bound structures (Table 1 and 2). Recent high-resolution structures of ABC transporter SBPs, P-type ATPases and Nramps are beginning to shed light on the key factors that drive Mn²⁺ specificity [31,75,81]. We summarize this progress below, as well as sequence analyses that provide clues regarding Mn²⁺ specificity for transporters like ZNT10 and ZIP8/14 [67].

Table 2.

Coordination geometry of metals in the binding site of manganese transporters

Family	Protein	Organism	Substrate	PDB code (Resolution in Å) [Reference]	Coordinating atoms (number)	Coordination geometry	RMSanglea (°)	Binding affinityb [Reference]c
Regulator	MntR	B. subtilis	Mn2+	1ON1 (1.75 Å) [84]	NO5 (6)	Octahedral	8	0.2 to 2 μM [83]
SBP of ABC transporter	PsaA	S. pneumoniae	Mn2+	3ZTTd (2.70 Å) [87]	N2O2 (4)	Octahedral (Tetrahedral)	20 (21)	3.3 ± 1.0 nM [87]
			Zn2+	1PSZ (2.00 Å) [74]	N2O2 (4)	Tetrahedral	13	231 ± 1.9 nM [87]
	SitA	S. pseudintermedius	Mn2+	4OXR (2.00 Å) [75]	N2O4 (6)	Octahedral	17	N.D.e [75]
			Zn2+	4OXQ (2.62 Å) [75]	N2O4 (6)	Octahedral	27	N.D.e [75]
P-type ATPase	SPCA1	Homo sapiens	Mn2+	7YAJ (3.16 Å) [91]	O5 (X)	Square pyramidal	9	0.07 ± 0.01 μMf [90]
			Ca2+	7YAH (3.12 Å) [91]	O6 (X)	Octahedral	13	0.13 ± 0.01 μMf [90]
Nramp transporter	MntH	D. radiodurans	Mn2+	8E60 (2.38 Å) [81]	SO5 (6)	Octahedral	25	190 ± 30 μM
			Cd2+	8E6M (2.4 8Å) [81]	SO5 (6)	Octahedral	34	55 ± 15 μM

^aThe angular deviation from ideal geometry, calculated as root mean squared deviation.

^bAffinities (K_d) were obtained using ITC unless otherwise specified.

^cReferences are listed only if they differ from those in the “PDB code" column.

^dIn this review, we interpret the geometry as octahedral based on bond length cut-offs and the authors interpret it as tetrahedral.

^eITC shows binding to both metals but due to fitting issues, exact K_d could not be determined, although reported to in nM range.

^fApparent affinities (K_m) were obtained using colorimetric ATPase assay.

MntR

We first set the stage by examining the structure of MntR, the bacterial transcription regulator of genes important for Mn²⁺ biology and homeostasis [43,82]. In contrast to transporters that interact with metal ions only transiently, MntR is effectively a switch regulated by high-affinity Mn²⁺ binding. MntR coordinates Mn²⁺ in a near-ideal octahedral geometry favored by Mn²⁺ (Figure 2A-C, Table 2) [83,84]. The metal-oxygen and metal-nitrogen coordinating bond lengths are within the ‘strong interaction’ range [78,85]. This near-ideal coordination of Mn²⁺ by MntR reflects its high affinity for Mn²⁺ (K_d of 0.2–2 μM) [83], and its function as a Mn²⁺-regulated switch.

Figure 2. Coordination chemistry impacts metal uptake in Mn²⁺ transporters.

(A) Structure of a Mn²⁺ transport regulator, MntR, from Bacillus subtilis, bound to Mn²⁺ (PDB ID: 1ON1; [84]). (B) The six Mn²⁺ ligands form a near-ideal octahedral coordination sphere. (C) Schematic of ideal geometry for octahedral, tetrahedral, and square pyramidal coordination spheres. (D) Structure of an SBP bound to Mn²⁺ (PDB ID: 3ZTT; [74,87]). (E-F) Metal-binding site of PsaA bound to Mn²⁺ (E; PDB ID: 3ZTT) and Zn²⁺ (F; PDB ID: 1PSZ). Each coordination sphere has a distinct geometry consistent with the ion’s preference: octahedral for Mn²⁺ (although two metal-oxygen bonds are longer, grey) and tetrahedral for Zn²⁺ (with the same two oxygens now more distant). The geometry is less distorted for Zn²⁺ than Mn²⁺ (for either interpretation of the structure as a the tetrahedral-like or octahedral-like sphere), consistent with the Zn²⁺ complex being more stable (Table 2). These differences allow Mn²⁺, the primary substrate, to be released by PsaA and transported, whereas the Zn²⁺ complex locks PsaA in a transport-incompetent closed conformation [31]. (G-H) Metal-binding site of SitA bound to Mn²⁺ (G; PDB ID: 4ORX) and Zn²⁺(H; PDB ID: 4OXQ) [75]. Both metals are coordinated octahedrally, less distorted for Mn²⁺ compared to Zn²⁺ (Table 2), making Mn²⁺ a preferred substrate [75]. (I) Structure of a P-type ATPase, human SPCA1 (PDB ID: 7YAJ), involved in Ca²⁺ and Mn²⁺ uptake into Golgi, bound to Mn²⁺ [91]. (J-K) Metal-binding site of SPCA1 bound to Mn²⁺ (J; PDB ID: 7YAJ) and Ca²⁺ (K; PDB ID: 7YAH) [91]. Mn²⁺ binds in a square pyramidal and Ca²⁺ in an octahedral geometry. Although Mn²⁺ binds in a less favored coordination, the Mn²⁺ coordination sphere has shorter bond lengths and less angular deviation compared to the Ca²⁺ coordination, probably resulting in tighter binding than Ca²⁺ (Table 2). (L) Structure of bacterial Nramp (PDB ID: 8E60), a divalent transition metal importer [34,136], bound to its physiological substrate, Mn²⁺ [81]. (M-N) Metal-binding site of DraNramp bound to Mn²⁺ (M; PDB ID: 8E60) and Cd²⁺(N; PDB ID: 8E6M) [81]. The coordination sphere of both metals comprises six ligands. D56—a key residue in metal and proton transport conserved across the Nramp family [76,247]—binds Mn²⁺ and not Cd²⁺. The coordination is more distorted from ideal octahedral geometry and the bond distances are longer in the Cd²⁺ complex than in the Mn²⁺ complex (Table 2). These differences indicate that Nramps, with a broad substrate profile, display plasticity in substrate binding to transport metal ion substrates with different properties. The angular deviations from ideal geometry, calculated as root mean squared deviations (RMS_angle), are indicated above each illustrated binding site.

SBPs

Structures of bacterial ABC transporter SBPs that import Mn²⁺—Streptococcus pneumoniae PsaA Staphylococcus pseudintermedius SitA—bound to different metals shed light on their metal-binding properties (Figure 2D) [31,75]. Both Mn²⁺ and Zn²⁺ bind at the same site in PsaA. The coordination geometry has been assigned as tetrahedral in both cases [31], although two additional oxygens at 2.4 Å—still within the ‘strong interaction’ range [78-80]—can be included to the Mn²⁺ sphere to produce an alternative distorted octahedral coordination (Figure 2E-F, Table 2). Mn²⁺ prefers octahedral and Zn²⁺, tetrahedral geometry [27,30]. Thus, Mn²⁺ binds PsaA in a non-ideal coordination geometry and with greater angular deviations than Zn²⁺ (Table 2). Based on the metal-free structure of the transporter domain, PsaC, its proposed metal-binding site includes two aspartates and two histidines, potentially forming a coordination sphere similar to its SBP, PsaA [86]. Interestingly, PsaABC transports Mn²⁺ and not Zn²⁺ although both metals bind PsaA with high affinity (K_d = 3.3 ± 1.0 nM for Mn²⁺ and K_d = 231 ± 1.9 nM for Zn²⁺; Table 2) [87]. These data suggest that the binding-site distortions may enable the conformational switching between different states of PsaA essential for Mn²⁺ transport. In contrast, the ideal geometry and higher thermal stability of the Zn²⁺-bound PsaA complex indicate reduced conformational flexibility [31]. Thus, Zn²⁺ appears to lock PsaA in a closed state and prevents transport [31]. In sum, the coordination geometries of the PsaA-metal complexes are key for selective transport of Mn²⁺ over Zn²⁺ by PsaABC, but also make Staphylococcus pneumoniae susceptible to Zn²⁺, which can inhibit the acquisition of the essential metal Mn²⁺ [31,88].

In the SitA case, Mn²⁺ and Zn²⁺ are both octahedrally coordinated with the same ligands, a coordination geometry preferred by Mn²⁺ but not Zn²⁺ (Figure 2G-H, Table 2) [75]. Moreover, some bonds to Zn²⁺ are longer than the optimal 2.0–2.2 Å and the coordination sphere is more distorted for Zn²⁺ than Mn²⁺ (Table 2). Although SitA binds both metals in the nM range (Table 2), the coordination geometries indicate that its binding site is better adapted to accommodate Mn²⁺. The MntC SBP also binds its preferred substrate, Mn²⁺, in an ideal octahedral coordination [89].

These SBP structures demonstrate that selective Mn²⁺ transport by bacterial ABC transporters is likely related to how Mn²⁺ and potential competing substrates are coordinated by the SBP. However, formation of an ideal coordination sphere does not necessarily result in effective transport. For example, effective transport of Mn²⁺ likely also depends on how readily the Mn²⁺ can be released from the SBP by the transmembrane subunits of the transporter, at least in the case of PsaA.

P-type ATPase

Human SPCA1, a P-type ATPase that exports both Ca²⁺ and Mn²⁺ from cytosol into the Golgi, with similar K_m values of 70 nM for Mn²⁺ and 130 nM for Ca²⁺ [90]. Cryo-EM structures of SPCA1 reveal how Ca²⁺ and Mn²⁺ bind with different geometry to the same site in the transmembrane domain (Figure 2I-K) [91]. Both metals bind with all oxygen ligands, with a 6-ligand octahedral geometry for Ca²⁺ and 5-ligand square pyramidal geometry for Mn²⁺ (Table 2), although in both cases, a potential water molecule ligand can be accommodated, which would generate a 7-ligand pentagonal bipyramid geometry commonly observed in Ca²⁺ coordination, and 6-ligand octahedral coordination for Mn²⁺. Bond lengths are smaller for Mn²⁺ than Ca²⁺ consistent with its smaller ionic radius. The small angular deviations and near-ideal bond length are consistent with the low K_m values for both substrates.

Nramps

In the bacterial Deinococcus radiodurans (Dra)Nramp, the physiological substrate Mn²⁺ and the non-essential substrate Cd²⁺ bind the same site, both in an octahedral geometry (Figure 2L-N) [81], although octahedral and tetrahedral geometries are the preferred coordination for Mn²⁺ and Cd²⁺ respectively [27,30]. Moreover, the angular deviations are higher for Cd²⁺ (Table 2). The coordinating ligands also differ, with different water arrangements and the aspartate sidechain (D56) in a different rotamer such that it coordinates Mn²⁺ but not Cd²⁺. This lack of coordination is consistent with Cd²⁺ being a softer metal than Mn²⁺, with lower propensity to bind harder oxygen ligands. This aspartate is essential for both metal and associated proton transport [34], and its interaction with Mn²⁺ may provide a mechanism to coordinate proton and metal transport. Interestingly, some coordinating bonds for Mn²⁺, and even more for Cd²⁺, are longer than ideal (Figure 2M-N). Metal-oxygen distances of 2.1-2.5 Å for Mn²⁺ and 2.3-2.8 Å for Cd²⁺ (0.2 Å longer for metal-sulfur) are considered ‘strong interactions.’ However, longer bonds of 2.6-3.2 Å for Mn²⁺ (and accordingly for Cd²⁺) are occasionally observed and considered ‘weak interactions’ still important for the overall coordination geometry [78].

DraNramp has relatively low affinity for both metals with K_d values in the ~100 μM range (Table 2; [81]). These are consistent with the observed non-ideal bond lengths and angles. The higher affinity for Cd²⁺ than Mn²⁺ is, at first glance, counterintuitive considering that Mn²⁺ is the physiological substrate and it binds with closer-to-ideal geometry. However, a lower affinity may lead to more favorable metal release kinetics to continue the transporter cycle. These data also suggest that information on substrate-binding affinity is useful, but often does not provide the full picture and additional factors are at play, like the energetics of the protein conformation.

Another interesting observation from the structures of DraNramp in different conformations is that the physiological substrate, Mn²⁺, binds at the same site but with different coordination geometries across the various biologically relevant conformations of its transport cycle (Figure 3) [81]. DraNramp is the only transporter for which Mn²⁺-bound structures are available in multiple conformations. All three (outward-open, occluded, and inward-open) conformations have octahedral Mn²⁺ coordination with substantial angular deviations. Interestingly, the bond lengths are longer and angular deviations are larger in the inward-open which is primed to release Mn²⁺ into the cytosol (Figure 3).

Figure 3. Coordination sphere changes across different conformations of an Nramp family Mn²⁺ transporter.

Structures of a bacterial Nramp (DraNramp) in three metal-bound conformations of the Mn²⁺ transport cycle—outward-open (PDB ID: 8E6N), occluded (PDB ID: 8E60) and inward-open (PDB ID: 8E6L)—shows that Mn²⁺ binds at the same site but with a unique coordination geometry in each of the three states [81]. Except for D56, N59 and M230, the bonding ligands vary across conformations along with the bond distances and the angular deviation (reported as RMS_angle). The coordination is most distorted in the inward-open conformation, which can facilitate Mn²⁺ release. These differences indicate that metal coordination can change not only for different substrates, but for different conformations of the protein bound to the same substrate.

CDF and ZIP transporters

While most CDF-family ZNT exporters selectively transport Zn²⁺, ZNT10 is implicated in Mn²⁺ transport in mammals [28,67]. A structure of Zn²⁺-bound human ZNT8 (Figure 4A), a Zn²⁺ transporter, confirms that the conserved HD-HD motif comprising two aspartates and two histidines forms the Zn²⁺-binding site (Figure 4B) [67,92]. In the absence of experimentally determined ZNT10 structures, sequence analysis suggests that an asparagine replaces one histidine to yield a ND-HD metal-binding motif [67]. This substitution may provide a hard oxygen ligand to bind the harder Mn²⁺ ion, and the asparagine could also form a bidentate interaction to expand its coordination sphere (Figure 4C).

Figure 4. Mn²⁺ binding by ZNT and ZIP transporters.

(A) Structure of human ZNT8 (PDB ID: 6XPE), bound to Zn²⁺ in different domains [92]. (B) Zn²⁺-binding site of the ZNT8 transmembrane domain with Zn²⁺ coordinated to two aspartates and two histidines in a tetrahedral geometry, a HD-HD motif conserved in most mammalian ZNTs [92]. (C) AlphaFold model of human ZNT10, a Mn²⁺ transporter, showing the residues analogous to the ZNT8 Zn²⁺-binding site with an asparagine (N43) replacing one of the histidines, now forming a ND-HD motif [67,248,249]. Asparagine likely provides a favorable oxygen ligand and a possibility for Mn²⁺ to expand its coordination geometry from tetrahedral to preferred octahedral in ZNT10 [67,92]. (D) Structure of BpZIP, a bacterial ZIP transporter (PDB ID: 5TSA), showing a bound Zn²⁺ at the transport site [93]. (E) Topology diagram of human LIV-1 subfamily ZIP transporters, showing that Zn²⁺ is transported between TM4 and TM5, which contains the conserved metalloprotease motif (HEXPHEXGD) [67]. (F) On TM5 of ZIP8 and ZIP14, the first histidine is replaced by an asparagine, which likely enables these transporters to transport Mn²⁺ in addition to Zn²⁺ [67,93,250]. In all panels, Zn²⁺ is shown as a blue and Mn²⁺ as a magenta sphere.

Among the mammalian ZIP family of Zn²⁺ importers, ZIP8 and ZIP14 can transport both Mn²⁺ and Zn²⁺ [67,93]. Structure of a bacterial ZIP transporter reveals the Zn²⁺ transport site (Figure 4D) [93], but in absence of human ZIP structures, sequence analyses provide clues about how ZIP8 and ZIP14 can transport Mn²⁺ [67,93]. Similarly, to the ZNT case above, in ZIP8 and ZIP14 a glutamate replaces the first histidine in the HEXPHEXGD motif conserved in TM5 of LIV-1-subfamily ZIPs (Figure 4E-F) [67]. This glutamate could provide one or two additional oxygen ligands to a Mn²⁺ coordination sphere.

Overview of iron transporters in different organisms

The key players in iron (Fe²⁺/Fe³⁺) homeostasis include transporters that transport iron in ionic or in complex form into and out of the cells and its compartments and receptors that bind and transport iron-bound complexes to their appropriate destinations (Figure 4, Table 3). Owing to the similarity in metal ion properties, there is a substantial overlap between iron and manganese transporters [28,64]. Since we have already introduced these overlapping transporters in the section on manganese transporters above, we will focus below on the proteins that are exclusively involved in iron homeostasis.

Table 3.

Transporters involved in homeostasis of iron

Family	Organism	Transporter	Function	Representative structurea
Nramp	Plants	NRAMPXb [10,47,137,138,174]	Fe2+ import into cytosol at PM, vacuole, and chloroplast	8E60c
	Fungid	Smf1/3 [2,34,136]	Fe2+ import into cytosol at PM (Smf1) and vacuole (Smf3)	8E60c
	Mammals	NRAMP1 [34,136,154,175]	Import of Fe2+ from phagosomes into cytosol	8E60c
	Mammals	NRAMP2 [34,136,154,175,176]	Fe2+ import into cytosol at PM, endosomes, and lysosomes	8E60c
ABC transporter	Gram-negative bacteriae	YbtPQ [177], YfeABCD [97,132,178], Irp6/7 [33,179], SfaABC [94,140,180], ViuPDGC [125,131], FhuCDB [131,181], FbpABC [27,182]	Fe2+ import and virulence	1Y4T
	Gram-positive bacteriae	Pit1/2 [183,184], IrtAB [185,186], PiaABC [131,187], FeuABC [124,131], MtsABC [133]	Complexed Fe2+ import and virulence	3HH8
	Plants	IDI7 [47,107,109]	Fe2+ import into cytosol from vacuolar tonoplast of leaves, stem, and flower	6G7P
	Plants	ATM3 [105,188,189] STA1 [106,190,191]	Fe2+ import into cytosol from mitochondria and biogenesis of Fe-S cluster	7N58f
	Fungid	Abc3 [110,192,193]	Heme import into cytosol from vacuole	3HH8
	Mammals	ABCB7 [120,154,188,194,195]	Complexed Fe2+ export from mitochondrial matrix	7VGFf
	Mammals	ABCG2 [120,196,197]	Complexed Fe2+ export at PM	6VXFf
CDF	Bacteria	YiiP [102], EmfA [103]	Fe2+ export	2QFIg
ZIP	Plants	IRT1/2 [47,105,109]	Fe2+ import at PM (IRT1 and IRT2) and vacuole (IRT2)	5TSBg
	Mammals	ZIP8, ZIP14 [15,17,119]	Import of free Fe2+ at PM	5TSBg
CCC1	Plants	VIT1 [49,50,138,198]	Fe2+ export from cytosol into vacuoles	6IU9
	Fungid	Ccc1, Pcl1 [2,110,113]	Fe2+ export from cytosol into vacuoles	6IU9
MFS transporter	Fungid	Arn1-4, Sit1, Enb1, Taf1, Str1/3 [2,94,110-112,199]	Internalization of Fe3+-siderophore (Arn1-4, Sit1, Enb1, Taf1, Str1) / Fe2+-heme (Str3)	5AYM
	Mammals	FLVCR1a/1b/2 [16,21,120,154,200]	Export out of (FLVCR1a) and import into (FLVCR1b/2) cytosol of heme bound Fe2+ at PM (FLVCR1a/2) and mitochondria (FLVCR1b)	5AYM
	Plants	FPN2 [105,108]	Fe2+ import into vacuole from cytosol	5AYM
	Mammals, bacteria	FPN [15,41,77,100,101,116,154,175,201]	Fe2+ export out of cytosol across PM and regulation of hepcidin	5AYM
	Mammals	HCP1 [21,22,94,202,203]	Import of Fe2+-heme into cytosol at PM	5AYM
Calcium channel	Mammals	LTCC [28,118,204,205]	Import of free Fe2+ into cytosol across PM	3JBRg
Porin	Gram-negative bacteria	Fe2+ porin [27,95]	Import of Fe2+ across OM	2PORh
Oligopeptide transportere	Plants	YSL [26,206]	Complexed-iron transport across chloroplast, vacuole, ER and PM	-
Permease	Plants	PIC1 [25,105]	Fe2+ export from cytosol into chloroplast	6C9Wg
	Fungid	Ftr1, Fip1 [2,94,110], Fet4 [2,94,110,207]	Fe3+ (Ftr1, Fip1) or Fe2+ (Fet4) import into cytosol across PM	6C9Wg
	Fungid	Fth1, Ftr2 [2,110]	Fe3+ import into cytosol from vacuole	6C9Wg
	Mammals	HRG1 [16,17,94,208,209]	Transport of Fe2+-heme into cytosol from PM, endosomes, and lysosomes	6C9Wg
Transferrin/Lactoferrin receptor	Gram-negative bacteria	TbpA/TbpB [94,210,211]	Import of Fe3+-transferrin into periplasm across OM	3V89
	Mammals	TfR1/2 [65,119,154,212], LfR [95,213-215]	Internalization of Fe3+-transferrin (TfR1/2), H-ferritin (TfR1/2) and lactoferrin (LfR)	3S9L
	Gram-negative bacteria	LbpA/LbpB [94,210,211]	Import of Fe3+-lactoferrin into periplasm across OM	7JRD
Heme receptor	Bacteriae	IsdABCH [94,126,216], HmuTUV [131,217], PhuTUV [127,131], TBUT heme receptors [218-221]	Import of Fe2+-heme into periplasm across OM	2Q8Q
	Fungid	Shu1 [110,192,222]	Import of Fe2+-heme into cytosol	2Q8Q
	Mammals	CD91 [17,223,224]	Import of Fe2+-heme in complex with hemopexin into cytosol across PM	2Q8Q
Hemoglobin receptor	Mammals	CD163 [17,223,225,226]	Import of hemoglobin bound Fe2+ in complex with haptoglobin into cytosol across PM	6K0Lg
Siderophore receptor	Bacteriae	Fe3+-siderophore receptors [123,131,227-230]	Internalization of Fe3+-siderophore across OM	4HMQ
	Mammals	24p3 [120,231,232]	Fe3+- siderophore import into cytosol across PM	3K3Lg
Ferritin receptor	Mammals	Tim-2 [120,154,233,234]	H-ferritin import into cytosol across PM	2OR7
	Mammals	Scara-5 [120,154,233,234]	L-ferritin import into cytosol across PM	7BZZ
	Mammals	SFT [154,173,235,236]	Tf-dependent and independent iron uptake into cytosol across PM in the intestine	-
G-protein coupled transporter	Gram-negative bacteria	FeoB [94,98,99,237-239]	Fe2+ import into cytosol across IM	3K53g
TonB	Gram-negative bacteria	TonB-ExbB-ExbD [94,96,240]	Energizes transport of iron-complexes from OM to IM (TonB)	2GSK
			Energizes transport of iron-complexes from OM to IM (ExbB-ExbD)	5ZFP
Mitochondrial carrier family transporter	Fungid	Mrs3/4 [2,110,114,115]	Fe2+ export from cytosol into mitochondria	1OKCg
	Mammals	MFRN1/2 [16,154,241-243]	Fe2+ export from cytosol into mitochondria	1OKCg
TRP channel	Mammals	TRPML1 [119,120,244,245]	Fe2+ export from endosomes and lysosomes into cytosol	5WJ5g

^aThe most relevant representative structure, prioritizing high-resolution structures with bound iron, then ones with other metal ions, then ones with no bound metal.

^bX is a number; plants generally have 5-7 numbered NRAMP homologs.

^cStructures with bound Mn²⁺.

^dFungi transporters are best characterized in yeasts, so yeast proteins are generally listed in this table.

^eNon-exhaustive list.

^fStructures with no bound divalent ions at sites for transition metal transport.

^gStructures with bound Zn²⁺.

^hStructures with bound Ca²⁺.

Bacteria

Iron is transported across the bacterial membrane mostly in complex form and sometimes in ionic Fe²⁺ form (Figure 5A, Table 3). In gram-negative bacteria, Fe³⁺-siderophores, Fe²⁺-heme, and Fe³⁺-transferrin iron is trapped at the OM by β-barrel protein receptors [27,29,94]. The TonB-ExbB-ExbD system energizes iron transport from the OM receptors across the periplasm to the inner membrane (IM) [27,95,96]. In both gram-negative and gram-positive bacteria, SBPs capture the complexed (siderophore with Fe³⁺ and heme with Fe²⁺) and ionic (transferrin with Fe³⁺) iron and their associated ABC transporters import iron into the cytosol [29,95,97]. Ionic Fe²⁺ in gram-negative bacteria is imported by OM porins and then by a GTPase, FeoB, across the IM [95,98,99]. Fe²⁺ export across the IM is mediated by the MFS-family transporter, FPN [41,100,101]; and the CDF-family YiiP and EmfA transporters [102,103].

Figure 5. Transporters and receptors involved in iron transport and homeostasis from bacteria to mammals.

(A) A gram-negative bacterial cell surface (left) highlighting outer membrane (OM) and inner membrane (IM) proteins involved in transporting Fe²⁺ and Fe³⁺ either as ions or as complexes (siderophore, transferrin, heme). The periplasmic-spanning complex TonB-ExbB-ExbD energizes transport of iron from the OM to the IM [27,94,96,240]. Most of the iron transport across the IM occurs through ABC transporters [27,29,94,95]. A G-protein coupled transporter, FeoB, enables Fe²⁺ import across IM [98,99,237,238]. Ferroportin (Fpn), an MFS family transporter, exports Fe²⁺ across the IM [41]. CDF transporters, YiiP and EmfA, are also implicated in Fe²⁺ export [102,103]. In gram-positive bacteria (right), iron import is mediated by PM ABC transporters with SBPs that directly capture iron ions or complexes [15,27,33,94]. The illustrated ABC transporters are a representative, but not exhaustive, list (see Table 3). (B) In a plant cell, iron ions are mainly transported across the cell membrane and organellar membranes by the transporters shown here [25,105,174,190]. (C) A representative fungal (yeast) cell highlighting proteins involved in ionic (Fe²⁺/Fe³⁺) and complexed (siderophore, heme) iron transport [2,94,110]. (D) In a typical mammalian cell, iron homeostasis is maintained by a combination of receptors—that bring iron in through endocytosis of heme, hemoglobin, transferrin, ferritin—and PM and organellar transporters [17,21,34,120]. (E) In a macrophage, iron is imported from phagosomes to the cytosol to deplete them of micronutrients as a strategy to kill engulfed pathogens, primarily by NRAMP1. A range of transporters and receptors also maintain the intracellular iron levels in macrophages [16,34,119,175]. In panels (C-E), some, but not all, oxidases and reductases that help in Fe²⁺-Fe³⁺ interconversion during transport are shown (see text). In each panel, specific proteins of each transporter family are shown, and their corresponding family is indicated in the legend at the bottom (see Table 3).

Plants

Previously introduced Mn²⁺ transporters also involved in iron homeostasis in plants include: YSL importers (PM, chloroplast, vacuole, ER) [25,104]; Nramp importers (PM, chloroplast, vacuole) [47,104,105]; importer IRT1 (PM) [24,105]; and VIT1 exporting Fe²⁺ into the vacuole [25,105]; (Figure 5B, Table 3). In addition, ABC transporter ATM3 exports iron from the mitochondria [24,105,106] and ABC transporter IDI7 imports Fe²⁺ from the vacuole into cytosol [47,107]; MFS-family FPN2 [105,108] and ZIP-family IRT2 [47,105,109] export Fe²⁺ from cytosol into vacuoles; and the PIC1 permease imports Fe²⁺ into the chloroplast stroma [25,105] (Figure 5B, Table 3).

Fungi

As with Mn²⁺, the fungal iron homeostasis machinery is best understood in yeasts (Figure 5C, Table 3). Key importers include: permeases Ftr1/Fip1 (PM) and Fth1/Ftr2 (vacuole) import Fe³⁺ [2,110]; permease Fet4 imports Fe²⁺ [2,94]; Nramp-family Smf1 (PM) and Smf3 (vacuole) import Fe²⁺ [2,34,94]; MFS-family Str3 (PM), heme receptor Shu1 (PM), ABC transporter Abc3 (vacuole), all import heme [110]; MFS-family Arn1-4, Sit1, Str1, Ernb1 Taf1 import Fe³⁺-siderophores across the PM [94,110-112].CCC1-family Ccc1 and Pcl1 export Fe²⁺ into vacuoles [2,110,113]. Finally, mitochondrial carrier (MC)-family Mrs3/4 import Fe²⁺ into the mitochondrial matrix [2,114,115].

Animals

Iron homeostasis in animals has many players that enable fluxes of free iron ions and iron complexes like Fe²⁺-heme and Fe-S clusters (Figure 5D and Table 3). Here we focus on proteins with major roles in transporting non-heme iron. In animals, iron primarily circulates as Fe³⁺-transferrin, which captured at the cell surface by TfR1/2 and internalized into endosomes [28,65,116]. Fe³⁺ from transferrin and other endocytosed complexes is released in acidic endosomes, converted to Fe²⁺ by reductases like STEAP3, and Fe²⁺ is then imported by NRAMP2 into the cytosol [21,34,116]. Cytosolic Fe²⁺ is either taken up in the mitochondria by MC-family MFRN1/2 for Fe-S cluster biogenesis and energy production, stored in ferritin for future use, or exported across the PM by ferroportin [17,101,117]. Additional Fe²⁺ importers include: ZIP8/14 (PM) [28]; the Ca²⁺ channel LTCC (PM) [28,118]; and the non-selective cation channel TRPML1 (endosomes and lysosome) [119,120].

Iron is central to the antimicrobial and erythrophagocytosis activities of macrophages. The iron transporters in macrophage cells are a subset of a regular animal cell, with the important addition of NRAMP1 at the phagosomal membrane that plays a key role in nutritional immunity, depriving phagocytosed pathogens of Fe²⁺ as it does for Mn²⁺ (Figure 5E) [34,121].

Structural insights into iron binding and transport

Iron is transported across membranes either as ions (Fe²⁺/Fe³⁺) or bound to heme or chelators like siderophores [65,122,123]. For these iron complexes, iron remains tightly bound to the respective chelator through the membrane transport process. In contrast, ionic forms of iron need to interact transiently with their transporters to enable rapid cycles of binding and release. Both Fe²⁺ and Fe³⁺ prefer octahedral geometry, with optimal bond lengths of 2.0-2.5 Å, although ‘weak interactions’ of 2.5-3.0 Å can still participate in a coordination sphere [78,80,85]. Here, we compare the coordination geometry and binding affinity of different iron binding sites, including ionic and complexed forms, to understand whether these properties help explain the distinct physiological functions of different transporters.

Siderophores, Heme and Transferrin

Structures of iron chelator complexes reveal near-ideal coordination geometry, consistent with their extremely high affinity for iron (Figure 6, Table 4). For example, the vibriobactin siderophore and heme (with contributions of the heme receptor protein PhuT) coordinate Fe³⁺ and Fe²⁺, respectively, octahedrally with optimal bond lengths and minimal angular deviation (Figure 6A-B, Table 4) [124-127]. Most siderophores are hexadentate and do not require additional interacting residues from the host proteins to complete Fe³⁺-octahedral coordination. An exception is tetradentate unsaturated siderophores captured by the periplasmic binding protein CeuE from Campylobacter jejuni. CeuE uses a histidine and a tyrosine to complete the near-ideal coordination of Fe³⁺ [128] (Table 4). Human transferrin—which mediates circulation of the poorly soluble Fe³⁺ in blood plasma while limiting levels of free Fe³⁺—binds Fe³⁺ with very high affinity (K_d ~ 10⁻²¹ M) [129]. Correspondingly, the Fe³⁺-transferrin also has near-perfect octahedral coordination geometry (Figure 6C, Table 4) [130]. According to the HSAB (hard and soft acid and base) theory, Fe³⁺ is a hard ion, and its ideal coordination sphere contains mostly hard oxygen ligands (with one nitrogen in both Fe³⁺ complexes illustrated in Figure 6). In contrast, Fe²⁺, a softer metal ion, and a heme provides four nitrogen ligands—with nitrogen being a softer ligand than oxygen.

Figure 6. Iron coordination in different receptors and transporters correlates to their biological function.

(A) Schematic representation of ideal geometry for octahedral and triagonal bipyramidal coordination spheres. (B-F) Structures showing Fe²⁺ or Fe³⁺ ions coordinated by six ligands in an octahedral geometry. The structures are: the Vibrio cholerae siderophore receptor ViuP with its vibriobactin siderophore bound to Fe³⁺ (B; PDB ID: 3R5T; [125]), the Pseudomonas aeruginosa heme receptor PhuT with a bound Fe²⁺-heme and residues Y71 and R228 completing the octahedral coordination (C; PDB ID: 2R79; [127]), human transferrin bound to Fe³⁺ (D; PDB ID: 3VE1; [130]), the metal-binding domain (MBD) of the Eucalyptus grandis vacuolar iron transporter 1 (VIT1) bound to Fe²⁺ (E; PDB ID: 6IU9; [50]), the SBP of a Yersinia pestis ABC transporter, Yfe1, bound to Fe³⁺ (F; PDB ID: 5UYE; [132]). (G) Bdellovibrio bacteriovorus ferroportin (BbFPN) bound to Fe²⁺ (PDB ID: 5AYM; [41]). Five ligands coordinate Fe²⁺ in a distorted triagonal bipyramid (Table 4; [41]). BbFPN is most distorted of all the iron coordination complexes (Table 4). The extent of distortion generally correlates with the biological function in that the chelators and binders need to bind iron with high affinity so that the metal is not released until the chelators reach their destination, whereas a lower affinity enables the transporters to transport metal readily. The angular deviations from ideal geometry, calculated as root mean squared deviations (RMS_angle), are indicated above each illustrated binding site.

Table 4.

Coordination geometry of iron in the binding site of transporters and receptors

Family	Protein	Organism	Substrate	PDB code (Resolution in Å) [Reference]	Coordinating atoms (number)	Coordination geometry	RMSanglea (°)	Binding affinityb [Reference]c
Siderophore receptor	ViuP	V. cholerae	Fe3+-vibriobactin	3R5T (1.45 Å) [125]	NO5 (6)	Octahedral	7	-
	FeuA	B. subtilis	Fe3+-enterobactin	2XUZ (1.90 Å) [124]	O6 (6)	Octahedral	7	~10−51 Md [246]
	CeuE	C. jejuni	Fe3+-catecholamide	5AD1 (1.32 Å) [128]	NO5 (6)	Octahedral	6	-
Heme receptor	PhuT	P. aeruginosa	Fe2+-heme	2R79 (2.40 Å) [127]	N5O (6)	Octahedral	6	-
Transferrin	hTf	H. sapiens	Fe3+	3VE1 (2.96 Å) [130]	NO5 (6)	Octahedral	14	~10−21 Me [129]
MBD of VIT transporter	VIT1	E. grandis	Fe2+	6IU9 (3.00 Å) [50]	SO5 (6)	Octahedral	7	1.9 ± 0.4 μMf [49]
SBP of ABC transporter	YfeA	Y. pestis	Fe3+	5UYE (2.09 Å) [132]	N2O4 (6)	Octahedral	22	17.8 ± 4.4 nM for Mn2+ [134]
	MtsA	S. pyogenes	Fe2+	3HH8 (1.87 Å) [133]	N2O4 (6)	Octahedral	22	4.3 μMg
Ferroportin transporter	BbFPN	B. bacteriovorus	Fe2+	5AYM (3.00 Å) [41]	O5 (5)	Triagonal Bipyramidal	30	195 μM for Co2+

^aThe angular deviation from ideal geometry, calculated as root mean squared deviation.

^bAffinities (K_d) are for iron (Fe²⁺ or Fe³⁺ as indicated in the “Substrate” column) and obtained using ITC, unless otherwise specified.

^cReferences are listed only if they differ from those in the “PDB code" column.

^dK_d obtained using visible spectroscopy.

^eK_d obtained using equilibrium dialysis studies.

^fK_d obtained using intrinsic protein fluorescence quenching.

^gK_d obtained using equilibrium dialysis and ICP-MS.

Metal-binding domains of transporters

VIT1 (vacuolar iron transporter 1), is primarily involved in iron export into vacuoles of plant and fungal cells, although it also transports other transition metals including Co²⁺, Zn²⁺, Mn²⁺ [49]. VIT1 has a cytosolic metal-binding domain (MBD) that is thought to capture iron before its eventual transfer to the transmembrane transport domain. A structure of the Eucalyptus grandis VIT1 MBD shows Fe²⁺ bound in a near-ideal octahedral coordination (Figure 6D, Table 4) [50]. However, the 4-Å bond to M149—a residue essential for metal transport—is longer than expected for a strong bonding interaction. Consistent with this observation, a K_d value of 1.9 ± 0.4 μM for Fe²⁺ binding to full-length VIT1 [49] suggests that none of its Fe²⁺-binding sites are as ideal as those of the chelators and binding proteins described above.

The SBPs of iron-specific ABC transporters also trap iron before transferring it to the transport module [131], and the iron binding geometry and affinity reveal a similar trend as that of VIT1. A representative SBP (YfeA, SBP of the iron importer YfeABC from Yersinia pestis) binds Fe³⁺ in its preferred octahedral coordination, but with angular distortions and some longer bond lengths (Figure 6E, Table 4) [132]. Again, similarly to VIT1, YfeA and MtsA (a Fe²⁺ SBP from Streptococcus pyogenes) have metal-binding affinities in the nM to low μM range (Table 4) [133,134].

Ferroportin

The bacterial ferroportin homolog Bdellovibrio bacteriovorus (Bb)FPN is the only protein for which we have a high-resolution structure with iron bound in the transmembrane transport domain. A structure of human ferroportin (FPN) bound to Co²⁺ provides clues of its Fe²⁺ binding and transport mechanism [135]. However, in absence of Fe²⁺-bound structures, we do not further discuss the human FPN structure. In BpFPN, Fe²⁺ binds in a non-ideal trigonal bipyramidal geometry with a high angular deviation and all coordinating bond lengths are longer than ideal (Figure 6F, Table 4). Moreover, BbFPN binds Co²⁺ (a transported substrate similar to, but more stable in solution, than Fe²⁺) with a K_d of 195 μM [41,77]. This affinity is similar to that observed for DraNramp and Mn²⁺. Mfrn1 from Oreochromis niloticus, another Fe²⁺ transporter, also display low affinity (K_d of 450 μM for Fe²⁺) [72]. Imperfect coordination geometry (bond angles and lengths) and poor binding affinity may be instrumental for ensuring conformational transitions and metal mobility required for transport.

Conclusions

The available Mn²⁺ transporter structures reveal that diverse coordination geometry and ligand type can enable Mn²⁺ binding and transport. Furthermore, distinct binding site structures result in different mechanisms for selectivity. Interestingly, in different Mn²⁺ transporters, either deviations from ideal bonding geometry or number of coordinating ligands allow both Mn²⁺ release and conformational changes seem to be required for transport to be kinetically accessible.

Similarly, analyses of the iron binders, chelators, and transporters show how both the ion binding properties and affinity correlate with the biological role of each type of iron interaction, with the binders (chelators and transferrin) having near-ideal coordination and much higher affinity towards iron compared to the transporters like BbFPN, and binding domains of transporters (SBPs and the VIT1 MBD), falling somewhere in between binders and canonical transporters.

Another general observation is that when comparing the binding of different metals to the same protein, affinity often poorly correlates with how ideal the coordination geometry is. This suggests that other energetic contributions are at play, such as conformational changes of the protein to accommodate the metal. Binding of a metal ion may cause a conformational change in the protein that incurs an energetic penalty, affecting the overall affinity of the protein for the metal. Furthermore, this conformational change may be on- or off-pathway in terms of the physiological function of the protein. For example, in DraNramp, the protein binds Mn²⁺ in a closer-to-ideal geometry than Cd²⁺ does, but that seems to come with some energetic cost because its affinity for Mn²⁺ is lower than that for Cd²⁺. In the case of PsaA, binding of Zn²⁺ is high affinity, but Zn²⁺ is not transported by PsaABC, presumably because the resulting PsaABC structure is not conducive to the next conformational steps in the transport cycle.

Overall, this review highlights how atomic-level knowledge of the coordination chemistry of different transition metals can help us understand the function of proteins in Mn²⁺ and iron homeostasis and how the transport behavior of the proteins may change during dyshomeostatic conditions.

Perspectives.

• The transition metals iron and manganese are essential to life, but toxic in high abundance. Metal ion homeostasis relies on a whole array of receptor, storage, regulatory, and transport proteins. Membrane transporters transiently interact with the metal ions to shuttle them into and out of cells and between cellular compartments, enabling both organismal and cellular metal homeostasis.

• Metal binding affinity and selectivity relies on the combination of how close the coordination geometry and ligands can approach ideal parameters, and how much conformational strain is induced in the protein to produce the resulting coordination geometry.

• There are still few examples of high-resolution structures of metal transporters in complex with metals. Additional structures are needed, as are studies of the metal binding thermodynamics and the protein conformational energy landscape, to fully understand and control the homeostatic machinery of manganese and iron.