Regulation of Mammalian Iron Metabolism and Biogenesis of Iron-Sulfur Proteins

In previous years, our laboratory identified and characterized the cis and trans elements mediating iron-dependent alterations in the abundance of ferritin and of the transferrin receptor. IREs are RNA stem-loops found in the 5′ end of ferritin mRNA and the 3′ end of transferrin-receptor mRNA. We cloned, expressed, and characterized the two essential iron-sensing proteins IRP1 and IRP2. IRPs bind to IREs when iron levels are depleted, resulting in either inhibition of translation of ferritin mRNA and of other transcripts that contain an IRE in the 5′ untranslated regions (UTR) or stabilization of the transferrin-receptor mRNA and possibly other transcripts that contain IREs in the 3′ UTR. The IRE–binding activity of IRP1 depends on the presence of an iron-sulfur cluster (see “Mammalian iron-sulfur–cluster biogenesis” below). IRP2 also binds to IREs in iron-depleted cells but, unlike IRP1, in iron-replete cells it is selectively ubiquitinated and then degraded by the proteasome.

To approach questions about the physiology of iron metabolism, we generated loss-of-function mutations of IRP1 and IRP2 in mice through homologous recombination in embryonic cell lines. In the absence of provocative stimuli, we initially observed no abnormalities in iron metabolism associated with loss of IRP1 function. Irp2^–/– mice develop a progressive neurologic syndrome characterized by gait abnormalities and axonal degeneration. Ferritin overexpression occurs in affected neurons and in protrusions of oligodendrocytes into the space created by axonal degeneration. Irp2^–/– animals develop iron-insufficiency anemia and erythropoietic protoporphyria. In animals that lack IRP1, IRP2 compensates for loss of IRP1's regulatory activity in most cell types, but we discovered several cell types and accompanying phenotypes in which Irp2 expression cannot be sufficiently increased to compensate. Animals that lack both IRP1 and IRP2 die as early embryos. The adult-onset neurodegeneration of adult Irp2^–/– mice is exacerbated when one copy of Irp1 is also deleted. Irp2^–/– mice offer a unique example of spontaneous adult-onset, slowly progressive neurodegeneration; analyses of gene expression and iron status at various stages of disease are ongoing. Dietary supplementation with Tempol mitigates neurodegeneration; the treatment appears to work by recruiting the IRE–binding activity of IRP1. We found that motor neurons were the most adversely affected neurons in Irp2^–/– mice and that neuronal degeneration accounted for the gait abnormalities. In a collaboration, we identified two IRP2^–/– patients who suffered from severe neurodegenerative disease in infancy and died before adolescence or were bed-ridden. A third infant with choreoathetosis established IRP2 deficiency as a cause of infantile and childhood neurodevelopmental disease.

We discovered that a transcript of the iron exporter ferroportin that lacks the IRE at its 5′ end is important in intestinal iron uptake. It allows ferroportin to permit iron to cross the duodenal mucosa in iron-deficient animals and also to prevent developing erythroid cells from retaining high amounts of iron. Our findings explain why microcytic anemia is usually the first physiological manifestation of iron deficiency in humans. Unexpectedly, we discovered that ferroportin is an abundant protein on mature red cells, where, as our work showed, it is needed to export free iron released from heme by oxidation. Using erythroid ferroportin-knockout animals, we showed that the absence of ferroportin results in accumulation of intracellular iron, increased oxidative stress, and reduced viability of cells in circulation.

We previously discovered that loss of IRP1 causes polycythemia and pulmonary hypertension by derepressing hypoxia-inducible factor 2-alpha (HIF2α) translation in the renal interstitium through the IRE–IRP system. Recently, we demonstrated that HIF2α expression is regulated mainly by IRP1 as the result of a bulge uridine in its IRE that interferes with binding by IRP2.

We also elucidated the pathophysiology of intravascular hemolysis and hyposplenism in animals that lack heme oxygenase 1 (HMOX1). Their tissue macrophages die because they cannot metabolize heme after phagocytosis of red cells. To mitigate or reverse disease, we performed bone-marrow transplants from wild-type animals to supply animals with functional macrophages, transplants that were successful. We then discovered that the transplant was not necessary by demonstrating that exogenously expanded wild-type macrophages can re-populate the reticuloendothelial system of Hmox1^–/– mice, restore normal erythrophagocytosis, and reverse renal iron overload and anemia. Five human HMOX1^–/– patients have been identified, but we believe that this represents an underdiagnosed and often misdiagnosed rare human disease. We are evaluating results from a large experiment on transcript expression in macrophages after red-cell phagocytosis.

Our goal in studying mammalian iron-sulfur biogenesis is to understand how iron-sulfur–prosthetic groups are assembled and delivered to target proteins in the various compartments of mammalian cells, including mitochondria, the cytosol, and the nucleus. We also seek to understand the role of iron-sulfur–cluster assembly in the regulation of mitochondrial iron homeostasis and in the pathogenesis of diseases such as Friedreich’s ataxia and sideroblastic anemia, which are both characterized by incorrect regulation of mitochondrial iron homeostasis.

The iron-sulfur protein IRP1 is related to mitochondrial aconitase, a citric acid–cycle enzyme, and it functions as a cytosolic aconitase in iron-replete cells. Regulation of the RNA–binding activity of IRP1 involves a transition from a form of IRP1 in which a [4Fe-4S] cluster is bound to a form that loses both iron and aconitase activity. The [4Fe-4S]–containing protein does not bind to IREs. Controlled degradation of the iron-sulfur cluster and mutagenesis reveal that the physiologically relevant form of the RNA–binding protein in iron-depleted cells is an apoprotein. The status of the cluster appears to determine whether IRP1 binds to RNA.

We identified numerous mammalian enzymes of the iron-sulfur–cluster assembly that are homologous to those encoded by the NIFS, ISCU, and NIFU genes, which are implicated in bacterial iron-sulfur–cluster assembly, and we observed that mutations in other iron-sulfur–cluster biogenesis proteins cause disease. We discovered more rare diseases that result from mutations in the mammalian biogenesis machinery, some of which predominantly affect cytosolic iron-sulfur biogenesis, such as CIAO1, the gene encoding cytosolic iron-sulfur–assembly component 1.

We identified a tripeptide motif, LYR, in many apoproteins that are recipients of nascent iron-sulfur clusters. The co-chaperone HSC20 (also known as HSCB) binds to HSPA9, its partner HSP70–type chaperone, and the chaperone complex binds to ISCU bearing a nascent iron-sulfur cluster that is delivered to iron-sulfur-cluster–recipient proteins. We identified several direct iron-sulfur–recipient proteins in a yeast two-hybrid assay, using HSC20 as bait. By studying one known iron-sulfur recipient, succinate dehydrogenase subunit B (SDHB), we discovered that several LYR motifs of the SDHB primary sequence engage the iron-sulfur–transfer apparatus by binding to the C-terminus of HSC20, facilitating delivery of the three iron-sulfur clusters of SDHB. We further discovered that the assembly factor SDHAF1 also engages the iron-sulfur–cluster transfer complex to facilitate transfer of iron-sulfur clusters to SDHB. The discovery of the LYR motif will aid in the identification of unknown iron-sulfur proteins, which are likely to be much more common in mammalian cells than had been previously appreciated. More recently, we discovered that, through recognition of LYR–like motifs in these recipient proteins, HSC20 is responsible for the delivery of iron-sulfur clusters to respiratory chain complexes I–II. Using informatics, we predicted that amino levulinic acid dehydratase (ALAD), a heme-biosynthetic enzyme, is a previously unrecognized iron-sulfur protein, and we identified more unrecognized iron-sulfur proteins by using the LYR motif to analyze candidate proteins. Using informatics, over-expression of candidate proteins, and iron detection with ICP–MS (inductively coupled mass spectrometry), we identified many more iron-sulfur proteins that are involved in a wide range of metabolic pathways, ranging from intermediary metabolism, DNA repair, and RNA synthesis, and possibly to the regulation of cellular growth. Iron-sulfur proteins will prove to be integral to the functioning and sensing of numerous pathways important in cellular functions.

Using informatics, we identified several potential iron-sulfur proteins encoded by SARS-CoV-2. We demonstrated that the SARs CoV-2 replicase Nsp12 ligates two cubane iron-sulfur clusters, one of which is needed for primer extension, whereas the other is needed for full assembly formation. The helicase encoded by the SARS-CoV-2 mRNA contains a cubane iron-sulfur cofactor in a domain known as the zinc-binding domain, which also ligates two zinc atoms. Using Tempol to degrade the iron-sulfur clusters, we stopped viral replication in tissue culture and greatly mitigated COVID19 disease in a Syrian Golden hamster model. Other coronaviruses also require iron-sulfur cofactors for function, including the original SARs and MERs, and likely, three causes of the common cold. We are actively pursuing studies and treatments for coronaviral disease based on these insights. We are studying a related coronavirus, OC43, which causes the common cold and can be studied safely. Our discovery that multiple iron-sulfur cofactors are present in the SARS-CoV-2 multimeric replication complex implies that viral exploitation of host iron-sulfur–biogenesis machinery and consumption of reducing equivalents is an important mechanism for viruses to tap into the energy stored in the cytosol of mammalian cells.