Nearly all organisms require iron to form the cofactors that activate essential enzymes driving cellular metabolism, small molecule synthesis, and organism self-renewal. In mammals, iron is the most abundant metal. The iron-containing heme cofactors found in oxygen-carrying proteins of red blood cells (hemoglobin) and muscle (myoglobin) are the most abundant metal-containing proteins in mammals (Dreyfuss et al., 1988). Iron deficiency is a clinical disorder that usually manifests as microcytic anemia with low levels of hemoglobin in blood. However, the essential requirement for iron cofactors extends beyond heme and the oxygen-carrying capacity of red blood cells.

Proteins coordinate iron cofactors in the form of heme, iron-sulfur clusters and mono- and di-nuclear iron centers. Heme is synthesized in every cell via an 8-step pathway that begins in the mitochondria, shifts to the cytosol, and ends in the mitochondrial inner membrane with the insertion of iron into protoporphyrin IX (Aasheim et al., 1994). Heme is poorly soluble in aqueous environments and has peroxidase activity that can cause oxidative damage. Nevertheless, heme enzymes locate to compartments throughout the cell; thus, heme needs to be safely transported across membranes and escorted through the cytosol to be inserted into its cognate enzymes.

Iron-sulfur clusters are assembled primarily in the mitochondrial matrix where they are inserted as [2Fe-2S] or [4Fe-4S] clusters into recipient apoenzymes of the mitochondria (Shi et al., 2008). An undefined sulfur-containing molecule is exported from the mitochondria via the transporter Atm1 (ABCB7) for use in cytosolic iron-sulfur cluster assembly. Although this exported compound has not been clearly identified, some studies propose that a [2Fe-2S] cluster coordinated by glutathione may be transported (Nandal et al., 2011; Leidgens et al., 2013). A parallel system of assembly factors is proposed to operate in the cytosol (Frey et al., 2014). The earliest steps in cytosolic Fe-S assembly are not well-defined, but both [2Fe-2S] and [4Fe-4S] clusters are required in the cytosol. Multiple scaffold/assembly proteins coordinate the [4Fe-4S] clusters and are required for their subsequent transfer to cytosolic and nuclear targeting factors (Frey et al., 2016). Iron-sulfur clusters can be sensitive to both oxidation and reduction and are not stable as uncoordinated, “free” molecules in solution. Therefore, both heme and iron-sulfur clusters require distribution systems that facilitate their transfer from biosynthetic/assembly machineries to recipient apoenzymes. Protein chaperones for heme and iron-sulfur clusters have been identified and the molecular details of their distribution systems, while incomplete, are gradually coming into focus. Simple mono-nuclear and dinuclear iron centers that activate the “non-heme” iron enzymes consist of Fe(II) and do not require dedicated assembly systems. They do require distribution systems, however, and recent studies reveal the roles that the iron chaperones play in maintaining cellular iron balance and utilization (Patel et al., 2019).

Iron ions are taken up by cells and initially deposited into the cytosol by iron-specific transporters, where they may be routed to different cellular destinations and used for various purposes (Patel et al., 2019). Most of the iron is used in mitochondria to synthesize heme and Fe-S clusters, to metallate mitochondrial non-heme iron enzymes, and, in some cells, to be stored in mitochondrial ferritin. Cytosolic iron is also used to form Fe-S clusters, to metalate non-heme iron enzymes, or oxidized for storage in cytosolic ferritin. Although these iron pools are often discussed as separate entities, they are highly dynamic and exchangeable. Cells exhibit a continuous, basal level of mitochondrial turnover and ferritin degradation in lysosomes, with recovered iron reentering the cytosol. Iron cofactor-containing enzymes have a limited half-life and are also continuously turned over and the iron recycled. Iron cofactors themselves may be damaged or lost, requiring repair or replacement, although less is known about these non-degradative processes. Finally cytosolic iron may be exported via the iron efflux pump, ferroportin. Ferritin may also be exported via exosomes or non-classical secretion, but the amount of iron exported as ferritin is low (Leffers et al., 1995). This highly dynamic pool of cytosolic iron has been referred to as the “labile iron pool” or LIP and is both kinetically exchangeable and chemically reactive. Theoretically, the cytosolic LIP is coordinated by a complex buffer consisting of large and small molecules. Empirical data indicate the major protein coordinating the LIP in the cytosol is poly C-binding protein 1 (PCBP1).

Studies of human physiology make clear that cells and tissues may have vastly different systems and requirements for handling and using iron. These differing tissue-specific patterns of iron homeostasis must also be integrated so that total body iron remains in balance, without overload or deficiency. Some cells and tissues are thought of as “professional” iron-handlers and exhibit high levels of iron transport, iron storage or iron utilization. Examination of the roles of PCBP iron chaperones in professional iron handling tissues has revealed molecular details of the functions of iron chaperones. Global deletion of the genes encoding PCBP1 or PCBP2 in the mouse results in lethality in early embryogenesis (Ghanem et al., 2016; Ryu et al., 2017). Tissue-specific, conditional deletion models have proven to be much more informative. Here, we describe approaches used within our laboratory to understand the roles of iron chaperones in cells and tissues.

Because developing erythrocytes require a tremendous uptake of iron to initiate and complete the heme synthesis needed for terminal differentiation, we hypothesized they were also likely to require iron chaperone activity. In a mouse model of inducible PCBP1 deletion, mice developed microcytic anemia that was traced to a defect in PCBP1-mediated ferritin iron storage (Ryu et al., 2017). In vitro and ex vivo murine models of proerythroblast development demonstrated that the earliest phases of terminal differentiation involved PCBP1-dependent storage of iron in ferritin. In PCBP1-depleted cells, this ferritin iron accumulation was impaired and associated with diminished heme and hemoglobin synthesis, which in the animal led to a microcytic anemia typical of iron deficiency. These studies relied on a new approach to studying erythroid development: an inducible erythroid differentiation system and the tracking of intracellular iron using radiotracers.

We used an estradiol-inducible murine cell line, G1E-ER4, that recapitulates terminal differentiation from the proerythroblast to orthochromatic erythroblast stage (Welch et al., 2004). The late stages of red blood cell development feature tremendous uptake of transferrin-bound iron, which is initially stored in ferritin and is subsequently transferred to mitochondria for insertion into heme. Heme is then incorporated into nascent globin chains to form hemoglobin. Radioisotopes of iron have long been used to measure uptake and absorption of iron in cells and animals. By labeling the developing erythroid cells with ⁵⁵Fe, we were able to track the flux of iron through 48 h of development: first in ferritin, then in heme, and finally in hemoglobin. We traced ⁵⁵Fe in native gels, which separate ferritin iron from hemoglobin iron, and measured the iron by phosphorimaging. ⁵⁵Fe in newly synthesized heme was organically extracted and measured by scintillation counting. Using RNA interference-mediated depletion of PCBP1, we measured the impaired delivery of ⁵⁵Fe to ferritin, the reduced synthesis of heme, and lowered levels of hemoglobin. Conversely, G1E-ER4 cells depleted of PCBP2 exhibited higher levels of ferritin storage and heme and hemoglobin synthesis, suggesting a competition between PCBP1 and PCBP2 for cytosolic iron. We then shifted to our mouse model of PCBP1 depletion, which relies on a tamoxifen-inducible Cre-loxP system with PCBP1 flanked by loxP sites and a transgene carrying the Cre recombinase-estrogen receptor (Cre-ER^T2) fusion protein (Ryu et al., 2017). Bone marrow from tamoxifen-treated animals exhibited lower levels of PCBP1 and lower levels of heme while peripheral blood analysis showed microcytic anemia. Finally, we isolated proerythroblasts from spleens of phenylhydrazine-treated animals lacking PCBP1 and traced their trafficking of ⁵⁵Fe through ex vivo differentiation. The impaired ferritin iron storage and subsequent poor hemoglobin synthesis initially observed in G1E-ER4 cells was again observed in the differentiating splenic proerythroblasts. By combining radiotracer studies with genetic models of PCBP1 depletion, we uncovered the critical role of PCBP1-mediated iron storage in ferritin as an important early step in red blood cell terminal differentiation.

Human adults require 1–2 mg of iron per day to compensate for the daily loss of iron (Leong et al., 2012; McKie et al., 2012). This requirement is met with dietary iron, which is primarily absorbed from the duodenum and proximal jejunum (Collins et al., 2018). Unlike other metals, iron cannot be excreted in humans. Thus, dietary iron absorption from the small intestine must be strictly regulated to ensure sufficient uptake to meet body iron requirements while simultaneously preventing oxidative damage to cells and tissues by excess iron accumulation. Both insufficient and excess iron absorption lead to a host of human health problems, such as iron deficiency anemia and hereditary hemochromatosis, respectively (Pasricha et al., 2021).

Dietary iron occurs in two forms, heme and non-heme iron. Studies in humans suggest that, while heme iron accounts for only about 10% of total dietary iron (Leong et al., 2012), it can comprise up to 40% of the total iron absorbed due to its bioavailability (Hurrell and Egli, 2010). However, the mechanism of heme iron absorption in animals is not well understood (Gulec et al., 2014) and in the mouse, nearly all dietary iron absorption occurs through non-heme iron (Dutt et al., 2022). Duodenal epithelial cells take up dietary iron through DMT1 at their apical membranes (Gunshin et al., 1997; Romero et al., 1997; Yanatori and Kishi, 2019). After dietary iron is imported into the enterocyte by DMT1, intracellular iron may be stored in ferritin, the iron storage protein, or be exported to the systemic circulation via the only iron exporter, ferroportin, which is located in basolateral membranes (Abboud and Haile, 2000; Donovan et al., 2000; McKie et al., 2000; Rishi et al., 2021). At the systemic level, intestinal iron absorption is regulated by hepcidin, a 25-amino acid-peptide hormone secreted by the liver and other tissues that binds ferroportin on basolateral membranes of intestinal epithelial cells and blocks its activity (Drakesmith et al., 2015; Nemeth and Ganz, 2021). If iron stored in enterocytes is not absorbed, it will be lost when the senescent epithelial cells are shed into the gut lumen. Thus, controlling iron efflux through ferroportin is the key regulatory step in dietary iron absorption. However, several studies have shown that ferroportin is not always responsive to the regulation of hepcidin (Frazer et al., 2017; Hanudel et al., 2022). Previous studies show that both PCBP1 and PCBP2 deliver iron to ferritin (Shi et al., 2008; Leidgens et al., 2013). In addition, PCBP2 has been shown to associate with DMT1 and ferroportin in yeast and mammalian cells (Yanatori et al., 2014; Yanatori et al., 2016). However, whether the iron chaperone activities of PCBP1 and PCBP2 play an important role in iron absorption through the small intestine or in maintenance of systemic iron balance has not been explored. There are still many questions that need to be answered before we can fully understand the mechanism of iron absorption through small intestine and intracellular iron trafficking in intestinal epithelial cells. Here we present some new approaches that became popular in the past decade and will be useful for future studies of iron absorption (Figure 2).

Radioactive isotopes ⁵⁵Fe and ⁵⁹Fe were widely used in iron absorption studies in vivo, both to determine mucosal iron uptake and to trace iron absorption from the small intestine to other organs. However, both ⁵⁵Fe and ⁵⁹Fe emit radioactive particles, which requires special laboratory equipment and skills, as well as specific handling and disposal of animals and samples treated with radioisotopes. While ⁵⁵Fe was initially used in human studies, the higher energy of ⁵⁹Fe rendered it inappropriate for human use. Stable isotopes of iron are becoming the preferred form of iron for biological and biophysical studies. There are four natural iron isotopes--⁵⁴Fe, ⁵⁶Fe, ⁵⁷Fe and ⁵⁸Fe, and their natural isotopic abundance ratios are 0.058, 0.917, 0.021 and 0.002, respectively (Taylor et al., 1992; Meija et al., 2016). The stable iron isotopes, ⁵⁷Fe and ⁵⁸Fe, are more commonly used for iron absorption studies in both human (Chen et al., 2005; Stoffel et al., 2017; Speich et al., 2021) and mouse (Fiorito et al., 2012; Nai et al., 2021) because of their low natural abundance. Historically, stable isotopes were not widely used in iron absorption studies due to their high cost and the limited availability of thermal ionization mass spectrometry (TIMS) for isotope ratio analysis (Chen et al., 2005). However, use of stable isotopes has increased with the development of high-resolution inductively-coupled plasma mass spectrometry (ICP-MS), which allows lower costs, easier sample preparation, and greater accessibility of isotope ratio analysis. ⁵⁴Fe, ⁵⁷Fe and ⁵⁸Fe allow studies on long-term and short-term iron absorption through direct measurement of these isotopic iron tracers from blood and other tissues. Because there are very little safety concerns, these studies using stable iron isotopes could cover various groups of study subjects, such as young children, adults, and pregnant women (Moretti et al., 2015; Cai et al., 2018; Delaney et al., 2020). As humans and mice share the key iron transporters and iron regulation mechanisms, many studies on iron absorption and iron homeostasis are carried out in mice, and more mouse studies use stable iron isotopes to trace iron trafficking and iron absorption (Fiorito et al., 2012; Chakrabarti et al., 2015; Sangkhae et al., 2020).

Tracer studies of iron absorption have become more powerful with the use of both genetic and pharmacologic methods to perturb iron trafficking. Minihepcidin is a synthetic peptide which contains the 9N-terminal amino acids of human hepcidin (Preza et al., 2011; Ramos et al., 2012). Based on mouse studies, minihepcidin can mimic the biological function of natural hepcidin and block iron export through ferroportin, making it a potential drug candidate for treatment or prevention of iron overload. Minihepcidin has been used in murine studies on iron absorption and iron homeostasis (Guggisberg et al., 2022; Hanudel et al., 2022) and can be used in conjunction with tracer studies to produce rapid and complete inhibition of ferroportin-mediated iron efflux.

In vitro studies of intestinal transport have relied in the past on immortalized cell lines cultured on membranes to mimic the epithelial layer of the intestine. Enteroids (intestinal organoids) are three-dimensional epithelial structures grown ex vivo from isolated intestinal stem cells (Sato et al., 2009). Human and mouse enteroid models are much more physiologically similar to mammalian intestine in vivo when compared to widely-used cultured cell models, such as Caco-2 and HT29 cells (Zachos et al., 2016; Rodrigues and Failla, 2021). Stem cells in undifferentiated enteroids can proliferate for many passages, allowing their propagation in culture. Under differentiation conditions, the stem cells can differentiate into all other cell lineages of intestinal epithelium, such as enterocytes, goblet cells, and enteroendocrine cells. Because of their capacity to recapitulate the major structural and functional characteristics of epithelium in vivo, enteroids are widely used in studies of cancer, gastrointestinal infection, and drug development (Zou et al., 2019; Schwartz et al., 2021). Enteroids were also used as models of small intestinal ion and dietary nutrient transport (Foulke-Abel et al., 2016; Pierson et al., 2018). The use of minihepcidin in enteroid models can recapitulate the systemic regulation of iron efflux by hepcidin in vivo. Compared to animal models, enteroids represent an environment that is both stable and readily probed with drugs and small molecules. Simply through switching on and off the systemic iron regulator, the enteroid model provides a powerful tool to study iron uptake and efflux, as well as intracellular iron trafficking through intestinal epithelial cells.

Although enteroid cultures are typically conducted on a scale too small for isotopic tracer studies, they are amenable to microscopic studies of iron trafficking. Physiologically, dietary iron that is taken up by enterocytes enters the cytosolic LIP. There, iron can be retained in the cytosol, transferred to ferritin for storage, or exported via ferroportin. Our previous studies indicate that PCBPs may serve as buffers for iron in the cytosolic LIP and control its trafficking. To track the LIP in enteroids, we have used recently developed fluorescent iron indicators. RhoNox-4, commercially available as FerroOrange, is a highly-sensitive fluorescent probe for intracellular labile iron based on an Fe(II)-mediated deoxygenation of a tertiary amine N-oxide (Hirayama, 2019; Hirayama et al., 2020). Compared to conventional fluorescent probes of Fe(II), such as calcein and PhenGreen-SK, RhoNox-4 selectively responds to Fe(II) with a direct fluorescent signal and without interference by other metals in living cells. The sensitivity of RhoNox-4 suggests that it may detect both “free” Fe(II) and iron bound by PCBP1, based on its kinetic lability and the dissociation constant (K_d) of PCBP1-bound iron. Use of this fluorescent iron indicator will help us to understand the roles of PCBPs as iron chaperones in iron absorption through measurement of the LIP in living enterocytes of enteroids (Figure 3).

The role of iron chaperones is especially crucial for liver function because the liver is the central iron storage organ in the body. The portal circulation directly connects the liver to the flow of dietary iron from the gut and the secretion of recycled red cell iron from splenic red pulp macrophages. Because liver cells are exposed to significant fluctuations of plasma iron, proper handling of the intracellular labile iron pool is critical for liver function and systemic iron homeostasis. Liver diseases such as non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH), and alcoholic liver disease (ALD) are associated with impaired iron regulation (Nelson et al., 2012; Britton et al., 2016; Ali et al., 2022). Animal studies in mice and rats also demonstrate that excess iron leads to oxidative stress, steatosis, and liver damage (Handa et al., 2016; Folgueras et al., 2018). We examined the role of the PCBP1 iron chaperone in the liver. We bred PCBP1 flox/flox mice carrying the Alb-Cre transgene, which mediates hepatocyte- and cholangiocyte-specific Cre recombinase expression. The hepatic loss of PCBP1 resulted in liver damage with steatosis (elevated liver triglycerides) and dying hepatocytes (elevated plasma alanine transaminase, ALT) (Protchenko et al., 2021). Several observations indicate that loss of iron chaperone results in liver iron toxicity. First, although PCBP1-depleted livers are iron-deficient, dietary iron supplementation did not improve liver parameters. In contrast, a low-iron diet prevented the development of steatosis. Second, iron-overloaded livers and PCBP1-deleted livers share similar patterns in their transcriptomes and lipidomes. Iron overload and PCBP1 deletion increased the expression of lipid biosynthesis genes and activated transcriptional pathways for the Nrf2-mediated oxidative stress response, ferroptosis signaling, and DNA damage response. Also, iron-overloaded and PCBP1-deficient livers produced similar profiles of oxidized lipid species, indicating intracellular iron toxicity.

In the lipid-rich environment of hepatic tissue, the accelerated lipid peroxidation due to the loss of the iron chaperone can affect many aspects of cellular metabolism. PCBP1-deficient livers exhibited depletion of mitochondrial lipids cardiolipin and coenzyme Q. These lipid changes contributed to mitochondrial damage and manifested in multiple signs of mitochondrial dysfunction in PCBP1-depleted livers (Jadhav et al., 2021). Diets supplemented with antioxidants vitamin E or coenzyme Q partially prevented the development of steatosis and liver damage, and coenzyme Q was required to prevent mitochondrial dysfunction.

In vitro visualization of changes in lipid oxidation and labile iron in primary cells from PCBP1-deficient livers is a powerful approach to real-time monitoring of the function of the iron chaperone and intracellular redox status. BODIPY C11 is a lipophilic, ratiometric fluorescent sensor used to quantify lipid peroxidation (Pap et al., 1999). Visualization of intracellular redox status in primary hepatocytes isolated from wild-type and PCBP1-deficient livers using BODIPY C11 confirmed accumulation of oxidized lipids upon loss of the iron chaperone. Hepatocytes deficient in PCBP1 have alterations in the labile iron pool which can be visualized using the fluorescent iron sensor FIP-1. FIP-1 reacts with kinetically-labile, chemically-reactive ferrous iron in cells (Aron et al., 2016). Intact FIP-1 has two FRET-producing fluorophores (fluorescein and Cy3) attached to a Fe(II)-cleavable endoperoxide bridge. Fe(II)-triggered cleavage of the linker separates fluorophores and eliminates FRET. Fluorescence of cleaved and non-cleaved forms of FIP-1 permits ratiometric measurement of labile iron. More iron-reactive fluorescence of FIP-1 was observed in hepatocytes isolated from PCBP1-deleted or iron-loaded wild-type livers. Even as PCBP1-deleted livers were iron-deficient, loss of the iron chaperone increased the chemical reactivity of intracellular iron (II). These studies confirmed the function of PCBP1 as an essential chaperone for iron to counteract the chemical reactivity of Fe(II) and prevent the formation of harmful reactive oxygen species.

The complex phenotype of PCBP1-depleted livers indicates that PCBP1 is integral to ferroptosis defense. Ferroptosis is a non-apoptotic, regulated cell death caused by iron-dependent lipid peroxidation (Stockwell, 2022). Ferroptosis manifestations vary in different cell lines and tissues but share common features, such as elevated levels of iron-mediated lipid peroxidation causing loss of membrane integrity, enhanced cell permeability, swollen mitochondria, and cell death (Stockwell et al., 2017; Jiang et al., 2021). Ferroptosis inducers include drugs that deplete intracellular glutathione or inhibit Gpx4, a glutathione peroxidase that converts lipid hydroperoxides to lipid alcohols (Forcina and Dixon, 2019). Ferroptosis inhibitors are mainly lipophilic antioxidants, such as vitamin E, liproxstatin-1 and coenzyme Q, or iron chelators, such as desferrioxamine. Glutathione is a critical co-factor for Gpx4 activity and also functions to coordinate intracellular iron (Hider et al., 2021). The complex of PCBP1 with iron and glutathione has an essential and yet unexplored role in ferroptosis defense to safeguard reactive, labile iron of the cytosolic LIP.

Liver-specific knockout of PCBP1 is a helpful model to express and study the multifunctional properties of PCBP1 in vivo. We introduced intact, wild type PCBP1 or variants with mutations in amino acid residues critical for the coordination of iron or oligonucleotides in each KH-domain of PCBP1 into adeno-associated virus serotype 8 (AAV) under the hepatocyte-specific promoter of thyroxine-binding globulin (TBG) (Patel et al., 2021). We then injected AAVs with PCBP1 variants into mice with PCBP1-deleted livers. The transduced variant protein demonstrated more expression in pericentral than in periportal areas of the liver, in agreement with previous studies (Bell et al., 2011). Despite spatial heterogenicity, the overall expression of transduced constructs in livers, as measured by immunoblotting, was the same or higher than endogenous protein. Transduction of the intact PCBP1 variant complemented the loss of endogenous PCBP1 in the liver and restored all examined parameters to the wild-type level. However, the transduction of mutant variants produced distinct phenotypes. Nucleic acid-binding activity was essential to prevent cell death. Transduction of the variant lacking RNA/DNA-binding but retaining iron binding activity in PCBP1-deleted livers was associated with persistence of elevated plasma ALT, indicating persistent liver damage and cell death. The iron-coordinating activity of PCBP1 was essential to prevent steatosis and DNA damage. Transduction with the variant lacking iron chaperone activity but retaining nucleic acid binding in PCBP1-deleted livers resulted in persistent steatosis and increased triglyceride accumulation. Surprisingly, the iron coordinating activity of PCBP1 was also essential in preventing DNA damage. PCBP1-depleted livers demonstrated increased TUNEL staining, a sign of DNA strand breaks and DNA damage. Livers expressing the RNA/DNA variant, which lacks iron binding, also exhibited higher TUNEL staining, indicating accumulation of damaged DNA. Thus, in vitro and in vivo approaches using HEK293 cells and mouse models, respectively, confirmed that iron chaperone activity is essential to prevent DNA damage. In contrast, the RNA/DNA-binding activity of PCBP1 is critical for viability (Patel et al., 2021) (Figure 4).

The essential nature of PCBP1 and PCBP2 proteins requires new approaches to studying the independent and overlapping functions of PCBP1 and PCBP2. AAV-mediated delivery of tissue-specific, codon-improved iCre-recombinase has advantages over transgenic mouse models with constitutive or inducible expression of conventional Cre-recombinase (Shimshek et al., 2002). For example, using AAV-mediated delivery of iCre in a ferroportin-floxed mouse strain using AAV9-Ple261(CLDN5)-iCre, ferroportin was depleted in the retina and brain vascular endothelial cells (Baumann et al., 2019). Transduction of AAV8-TBG-iCre in JMJD3 fl/fl mice produced successful depletion of the iron enzyme histone lysine demethylase in liver (Seok et al., 2018). The use of AAV-iCre also makes possible the simultaneous deletion of a floxed gene and the expression of its mutant variant. This approach was used to co-transduce iCre and wild-type or phosphorylation-resistant Wwtr1 mutant into Wwtr1 fl/fl livers (Wang et al., 2020). Simultaneous deletion of essential PCBP1 and PCBP2 during embryogenesis might be too detrimental for hepatocytes. The AAVs expressing iCre recombinase from liver-specific TBG promoters offer a unique opportunity for coordinated deletion of PCBP1, PCBP2, or both.

While animal models are very useful for studying iron chaperones in a physiological context of iron trafficking, other approaches are needed to identify new components of iron cofactor trafficking and delivery systems. Unbiased genome- and proteome-wide screens to explore protein interactions are powerful tools for the identification of new proteins participating in intracellular iron trafficking pathways. Three main approaches have been used: forward genetic screens, affinity purification, and proximity labeling. PCBP1 was identified as the iron chaperone for ferritin in a forward genetic screen conducted in baker’s yeast. In this approach, a human liver cDNA library was screened for genes that could bind and facilitate iron delivery to human ferritins when both were exogenously expressed in yeast (Shi et al., 2008). Other forward genetic screens based on the classical yeast two-hybrid assay have been used to identify human proteins interacting the iron transporters DMT1 and ferroportin 1. These screens identified PCBP2 as a chaperone involved in facilitating iron uptake and efflux in HEp-2 cells (Yanatori et al., 2014; Yanatori et al., 2016). Approaches based on the use of mass spectrometry for the identification of interacting proteins involved in iron trafficking have proven successful. In general, the protein of interest (the “bait”) is used to capture the interacting proteins (the “prey”), which are purified, identified, and quantified using spectrometric and proteomic analyses. The two main approaches used to capture those interacting proteins are affinity purification and proximity labeling, which sometimes are used in combination (Liu et al., 2020).

Affinity purifications are based on the isolation of protein complexes derived from live cells. The bait protein is pulled down from a cell lysate using specific antibodies and the interacting proteins are co-purified with the bait. All proteins are then identified via mass spectrometry. Although the endogenous bait protein may be immunoprecipitated using a very high affinity antibody against the bait, the most common approach is to use an epitope-tagged version of the bait protein instead, as this lessens the chance that the antibody will interfere with the native protein complexes. Because overexpression of the bait protein can lead to non-native interactions, we choose to use stable cell lines that inducibly express FLAG epitope-tagged bait proteins at moderate levels. We immunoprecipitate under native conditions using anti-FLAG-agarose and perform mass spectrometry identification of the affinity-purified interacting proteins. Cell lines that do not express the FLAG-tagged bait are used in parallel to identify non-specific interacting proteins. This approach allowed us to identify BolA2 as a PCBP1-interacting protein; subsequent studies determined that that PCBP1 and BolA2 form an iron chaperone complex for the assembly of cytosolic [2Fe–2S] clusters (Patel et al., 2019). A similar approach has been used to identify key factors involved in Fe-S cluster assembly and distribution in the cytoplasmic Fe-S biogenesis pathway (CIA) (Gari et al., 2012; Stehling et al., 2012; Maio et al., 2021). The coimmunoprecipitation-mass spectrometry approach allowed the identification of a [4Fe-4S] cluster targeting complex for nuclear proteins that contained MMS19, CIAO1, and FAM96B. These studies also identified the chromokinesin KIF4A as a MMS19-interactor and facilitated the characterization of KIF4A as [4Fe-4S] cluster binding protein (Ben-Shimon et al., 2018). More recently, Maio and colleagues described the SARS-CoV-2 RNA-dependent RNA polymerase, Nsp12, as an Fe-S cluster protein and used coimmunoprecipitation with mass spectrometry analyses to demonstrate the interaction of Nsp12 with components of the Fe-S biogenesis pathways as well as PCBP1 and BolA2 (Maio et al., 2021).

Although multiple studies have successfully used coimmunoprecipitation and mass spectrometry to identify binding partners in iron trafficking pathways, the transient and weak nature of these interactions may prevent the co-purification of protein complexes. Further challenges arise when the buffer conditions needed to solubilize, e.g., the nuclear compartment, prove too harsh to maintain interactions. Detection may be improved with the inclusion of crosslinking agents such as formaldehyde, which trap associations between binding partners during the purification steps. Additionally, a recently developed, targeted mass spectrometry approach provides better sensitivity and quantitation and is applicable when the interacting partners are known or suspected. This approach was used to confirm that the cytosolic Fe-S cluster assembly complex forms a metabolon consisting of scaffolding components, the targeting complex, and recipient apo-proteins (Fan et al., 2022).

Proximity labeling constitutes a powerful alternative to detect transient, less stable protein interactions in living cells, as this approach can identify proteins that are briefly close to the bait protein. The most common approach is proximity-dependent biotin identification (BioID), which is based on the expression of the bait protein fused with a promiscuous prokaryotic biotin ligase, BirA* (Choi-Rhee et al., 2004; Roux et al., 2012). BioID fusion proteins biotinylate adjacent proteins, within 10–20 nm from bait, then streptavidin resin selectively captures the biotinylated proteins, which are identified and quantified using mass spectrometry. BioID has its own intrinsic advantages and limitations. The two main advantages of BioID are the detection of both transient and stable protein-protein interactions in living cells and the insensitivity to non-native interactions that may occur during coimmunoprecipitation. The main drawbacks are interference of the BirA* moiety with natural interactions and false positives that may arise from close proximity to irrelevant proteins (Roux et al., 2018). The simultaneous use of both coimmunoprecipitation and BioID approaches is complementary and can be achieved by using a single stable cell line construct and a multi-tag workflow (Liu et al., 2020).

All protein-protein interactions identified via proteomic analysis must be validated by other methodologies. These may include reciprocal coimmunoprecipitation or streptavidin pull-downs followed by immunoblotting (Frey et al., 2016). Protein pairs in close proximity (<40 nm) may be detected in situ by the use of proximity ligation assay (PLA). This method employs pairs of PLA probes, which are specific antibodies conjugated to complimentary DNA oligonucleotides. If the two tested proteins are close, PLA pairs can hybridize to make circular DNA, which can be amplified and fluorescently labeled for visualization by fluorescence microscopy (Alam, 2018).

While the biophysical and biochemical characteristics of other nutrient metal cofactors differ from those of iron cofactors, the transient nature of the interactions between metallochaperones and client proteins is also characteristic of other metal distribution systems, making the methods described here suitable for the analysis of other metal trafficking systems. For example, AAV transduction of murine models of metallochaperone deficiency is a versatile approach and has been used to study truncated forms of ATP7B in a mouse model of Wilson disease. Fluorescent metal sensors have been developed for Cu, Zn, Ca, Mg, etc., which are commercially available. For example, some protein interactors of copper chaperones, transporters and regulators have been uncovered by conventional yeast two-hybrid analysis (van Dongen et al., 2004) (Ohrvik and Wittung-Stafshede, 2015) (Lim et al., 2006a) (Lim et al., 2006b) (Larin et al., 1999) and coimmunoprecipitation followed by mass spectrometry analysis (Matson Dzebo et al., 2018) (Comstra et al., 2017). Examples of the use of BioID coupled to whole proteome analysis for the study of protein interactomes in the context of metal trafficking are scarce. The few examples include the use of BioID constructs to explore some protein interactors of the copper chaperone for superoxide dismutase (CCS) followed by detection via western blot without mass spectrometry (Grasso et al., 2021), and the interactome of the copper chaperone SCO1 determined during the description of a high-throughput BioID technique (Liu et al., 2018). Therefore, the use of the unbiased BioID approach described here for the study of the protein-protein interactions that drive the flux of metals inside cells is promising and remain to be explored. In addition to Fe-S clusters, cysteine residues coordinate other metals such as copper (I) and zinc (II) (Arnesano et al., 2002). Just as Fe-S clusters protect coordinating cysteine residues from alkylation (Bak and Weerapana, 2023), copper (I) and zinc (II) binding also attenuate the reactivity of the coordinating cysteine thiolates (Apuy et al., 2001; Glauninger et al., 2018; Novoa-Aponte et al., 2019). Hence, the cysteine reactivity assay seems feasible for the whole proteome analysis of apo and holo states of proteins in the context of copper and zinc trafficking, although this has yet to be demonstrated.