: S. Taketani, Department of Biotechnology, Kyoto Institute of Technology, Kyoto 606–8585 Japan. Fax: + 81 75 724 7760, Tel.: + 81 75 724 7789, E-mail: email@example.com
Mammalian ferrochelatase, the terminal enzyme of the heme biosynthetic pathway, catalyzes the insertion of a ferrous ion into protoporphyrin and contains a labile [2Fe−2S] cluster center at the C-terminus. To clarify the roles of the iron–sulfur cluster in the expression of mammalian ferrochelatase, enzyme activity in human erythroleukemia K562 cells under iron-depleted conditions was examined. Treatment of cells with an iron chelator, desferrioxamine, resulted in a decrease in enzyme activity, in a dose- and time-dependent manner. Heme content decreased during desferrioxamine treatment of the cells. Addition of ferric ion-nitrilotriacetate [Fe (III)NTA] to desferrioxamine-containing cultures led to restoration of the reduction in the enzyme activity. While RNA blots showed that the amount of ferrochelatase mRNA remained unchanged during these treatments, the amount of ferrochelatase decreased with a concomitant decrease in enzyme activity. When full-length human ferrochelatase was expressed in Cos7 cells, the activity was found mainly in the mitochondria and was decreased markedly by treatment with desferrioxamine. The activity in Cos7 cells expressing human ferrochelatase in cytoplasm decreased with desferrioxamine, but to a lesser extent. When Escherichia coli ferrochelatase, which lacks the iron–sulfur cluster, was expressed in Cos7 cells, the activity did not change following any treatment. Conversely, the addition of Fe (III)NTA to the culture of K562 and Cos7 cells led to an increase in ferrochelatase activity. These results indicate that the expression of mammalian ferrochelatase is regulated by intracellular iron levels, via the iron–sulfur cluster center at the C-terminus, and this contributes to the regulation of the biosynthesis of heme at the terminal step.
As the terminal enzyme of the heme biosynthetic pathway, ferrochelatase catalyzes the insertion of a ferrous ion into protoporphyrin IX to form protoheme, and the animal enzyme is located at the inner membrane of the mitochondria . Both cDNA and genes for ferrochelatase from various species including mouse, human, yeast, plant, Escherichia coli and Bacillus subtilis have been isolated [2,3]. The mammalian enzyme is nuclear encoded, as a precursor form (48 kDa) and translocated into the mitochondrion where the enzyme is processed proteolytically to its mature size of 41–42 kDa . The deduced amino-acid sequences from various species show 10–88% identity to the human enzyme [2,3]. Heterologous overexpression of ferrochelatase demonstrated that a conserved and essential histidine residue is involved in the binding of the metal substrate  and a conserved glutamate residue may play a role in the catalytic reaction . The mammalian enzyme contains a [2Fe−2S] cluster in the C-terminal region. However, yeast and plant ferrochelatase do not contain this motif, or alternatively the C-terminus of prokaryotic ferrochelatase is ≈ 30 amino acids shorter than that of the mammalian enzyme [6,7]. Furthermore, C-terminal-truncated human ferrochelatase has been reported to be inactive . Thus, the Fe/S cluster is essential for enzyme activity although the cluster is distinct from the binding region of the ferrous ions substrate . One approach to demonstrate the role of the cluster has been established; namely, the intracellular targets of NO gas correspond to various proteins containing an Fe/S cluster . Accordingly, NO inhibited purified recombinant human ferrochelatase as well as its activity in rat hepatocytes  and macrophage RAW264.7 cells . ESR spectroscopy provided evidence that NO interacts with the Fe/S cluster, while ferrochelatase activities of yeast and bacteria were not inhibited by NO . However, the essential roles of the Fe/S cluster in mammalian ferrochelatase remain unknown.
Iron is not only a substrate of heme biosynthesis but is also utilized in the formation of the Fe/S cluster of nonheme iron protein in the mitochondria, where the main prosthetic group of enzyme subunits for respiratory chain complexes contains an Fe/S cluster. Recently, three gene products of yeast including frataxin, an ABC transporter, Atm1p and cysteine desulfurase homolog Nfs 1p, have been shown to promote mitochondrial iron metabolism and be involved in the biosynthesis of Fe/S-containing proteins . Although it has been shown that iron import into yeast mitochondria to utilize a substrate for ferrochelatase was dependent on a membrane potential as an energy source , it is unclear whether the iron transporter supplying ferrochelatase is also involved in importing iron into the mitochondrial matrix for biogenesis of the Fe/S cluster. Moreover, interaction of heme synthesis with the biogenesis of proteins containing the Fe/S cluster in mammalian cells is unclear. Here, we report that the treatment of cells with iron-chelators resulted in a decrease in ferrochelatase activity in human erythroleukemia cells and also Cos7 cells expressing human ferrochelatase. The importance of the Fe/S cluster of ferrochelatase in the iron-dependent regulation of heme biosynthesis is the focus of this study.
Materials and methods
[α-32P]dCTP and Nylon membranes were purchased from Amersham-Pharmacia Co. Restriction endonucleases and DNA-modifying enzymes were from Takara Co. and Toyobo Co. Mesoporphyrin IX and coproporphyrin III were products of Porphyrin Products. Desferrioxamine was obtained from Sigma Chemical Co. Coproporphyrinogen was prepared from coproporphyrin by reduction with sodium amalgam. Antibodies against ferrochelatase and coproporphyrinogen oxidase were as described previously . cDNA probes of human ferrochelatase and coproporphyrinogen oxidase were as described previously [15,16]. All other chemicals were of analytical grade.
Human erythroleukemia K562 cells were grown in RPMI-1640 medium supplemented with 10% fetal bovine serum and antibiotics. Cos7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 7% fetal bovine serum and antibiotics. Cos7 cells were transfected with 2 µg of plasmid, using a transfection reagent Lipofectamine plus for 4–5 h, according to the manufacturer's recommendations. Cells were then incubated in DMEM supplemented with 10% fetal bovine serum for 5 h and further incubated in the absence or presence of desferrioxamine for 16 h. Heme and protoporphyrin contents of the cells were determined as described previously [14,17].
The expression vector pCD , linearized by digestion with EcoRI, was ligated with whole insert derived from human ferrochelatase cDNA, HF2.1  and the resulting plasmid was designated pCD-HF. To make pCD-HFΔL, which lacks part of the first 53 amino acids of the N-terminus corresponding to the leader peptide of the human ferrochelatase [15,19], two primers, 5′-TTGAATTCATGGGTGCAAACCTTCAAGTT-3′ and 5′-ATGAATTCTCACAGCTCCTGGCTGGT-3′ were synthesized. pAR-HF, which encodes the human ferrochelatase in the E. coli expression system and the product that lacks the region for a putative leader peptide of ferrochelatase , was digested with EcoRI and amplified via PCR using the resulting plasmid as a template. After the product had been digested with EcoRI, the DNA was ligated into the EcoRI-digested pCD vector. To construct a plasmid carrying E. coli ferrochelatase, two primers, 5′-TTGAATTCATGCGTCAGACTAAAACC-3′ and 5′-AAGAATTCTCAGCTTTAGCGGGCAA-3′, were synthesized. PCR was carried out with SalI-digested pFC3 encoding E. coli ferrochelatase  as a template. The product was digested with EcoRI and ligated into EcoRI-digested pCD vector. The nucleotide sequence of the constructs was verified by DNA sequencing.
Total RNA was isolated from K562 cells using the guanidium isothiocyanate method . Twenty micrograms of RNA was applied to a 1% agarose gel and electrophoresed, subsequently transferred onto a nylon membrane (Amersham Hybond N+) for hybridization with DNA probes. The filters were hybridized, washed and exposed to an X-ray film as described previously .
The lysates of K562 cells and Cos7 cells untreated or treated with desferrioxamine were subjected to SDS/PAGE and electroblotted onto poly(vinylidene difluoride) (PVDF) membrane. Immunoblotting was carried out with antiferrochelatase or anticoproporphyrinogen oxidase antibody as the primary antibody .
Ferrochelatase activity was measured using mesoporphyrin and zinc acetate as substrates and the activity of coproporphyrinogen oxidase was determined as described previously [14,16]. To isolate the mitochondria-rich fraction of Cos7 cells, cells were suspended with 10 mm Tris/HCl buffer, pH 8.0, containing 80 mm KCl and 0.25 m sucrose and homogenized with 10 strokes with a Teflon pestle. Unbroken cells and cell debris were removed by centrifugation at 500 g for 5 min at 4 °C. Homogenates were then centrifuged at 12 000 g for 10 min at 4 °C, and the resulting pellet was washed twice with the above buffer. The thus obtained 12 000 g pellets as the mitochondrial fraction and the 12 000 g supernatant as the post mitochondrial supernatants were used for the enzyme assay, respectively. Protein concentration was estimated by the Bradford method .
Changes in the expression of ferrochelatase and heme content in K562 cells by treatment with desferrioxamine
To examine whether ferrochelatase activity is altered by the change in the intracellular iron level, human erythroleukemia K562 cells were treated with an iron chelator, desferrioxamine and the activity was measured. Ferrochelatase activity in cells with 50 µm desferrioxamine for 16 h decreased to 80% of control, while that with 100–150 µm desferrioxamine decreased to 44–50% (Fig. 1A). The presence of 160 µm Fe (III)NTA and 150 µm desferrioxamine restored the activity. The activity of coproporphyrinogen oxidase, the expression of which may be independent of iron metabolism, was also examined as the control. Enzyme activity in desferrioxamine-treated cells remained unchanged. Ferrochelatase activity decreased in a time-dependent manner. Another iron chelator, sodium 4,5-dihydroxybenzene-1,3-disulfonate (Tiron) at 1 mm also decreased ferrochelatase activity (Fig. 1B). RNA blots showed that mRNAs for ferrochelatase and coproporphyrinogen oxidase in desferrioxamine-treated cells remained unchanged, compared with the control cells (Fig. 2A). In contrast, the amounts of ferrochelatase in desferrioxamine-treated cells decreased, accompanied by a decrease in the enzyme activity, whereas those of coproporphyrinogen oxidase remained unchanged (Fig. 2B). These results suggest that the decrease in intracellular iron levels results in a decrease in ferrochelatase level, possibly due to proteolytic degradation of the ferrochelatase protein. To examine whether the change in ferrochelatase level affects heme biosynthesis, the total content of heme and protoporphyrin in K562 cells treated with desferrioxamine for 16 h was measured. Heme content in cells tended to decrease in the presence of > 100 µm desferrioxamine, while protoporphyrin content remained unchanged under these conditions (Fig. 3).
Possible requirement of the Fe/S cluster for iron-dependent regulation of the expression of ferrochelatase in Cos7 cells
The above results indicated that the regulation of ferrochelatase by iron is at the post-translational level. To clarify the role of the Fe/S cluster at the C-terminal region of mammalian ferrochelatase, human and bacterial ferrochelatase were transfected into Cos7 cells. Furthermore, to examine whether transiently expressed human ferrochelatase is located in mitochondria, cells were fractionated into mitochondrial and post-mitochondrial supernatant fractions and enzyme activity were measured. Enzyme activity in human ferrochelatase (pCD-HF)-transfected cells was found mainly in the mitochondrial fraction (Fig. 4A). When human ferrochelatase lacking the first 53 amino acids of the N-terminus, corresponding to the leader peptide of the enzyme (pCD-HFΔL), was expressed in Cos7 cells, enzyme activity in the post-mitochondrial supernatant corresponding to cytosol was higher than that in the mitochondrial fraction (Fig. 4A). Transfection of E. coli ferrochelatase (pCD-EF) also resulted in cytosolic expression of the enzyme. Figure 4B shows the effect of desferrioxamine on ferrochelatase activity in Cos7 cells expressing human and E. coli ferrochelatase. Treatment of pCD-HF-transfected Cos7 cells with 50–150 µm desferrioxamine led to decreases to 42–71% of the ferrochelatase activity of the control and this decrease was restored by addition of 160 µm Fe (III)NTA. Enzyme activity in pCD-HFΔL-transfected cells decreased following treatment with desferrioxamine, but to lesser extent. When pCD-EF-transfected cells were treated with desferrioxamine, the activity mostly remained unchanged. Immunoblot analysis revealed that the amounts of ferrochelatase protein in pCD-HF-transfected cells was decreased markedly by treatment with 100 µm desferrioxamine (Fig. 4C). The level of the precursor form of ferrochelatase in desferrioxamine-treated cells was similar to that in control cells. These results suggest that the Fe/S cluster and mitochondrial location of ferrochelatase may be required for regulation of the expression of mammalian ferrochelatase by intracellular iron levels.
Iron positively regulates the expression of ferrochelatase
Finally, we examined the additive effect of iron on ferrochelatase expression. When pCD-HF-transfected Cos7 cells were treated with Fe (III)NTA, ferrochelatase activity in the cells increased and reached 1.5-fold, up to 100 µm, followed by a decrease at increased iron concentrations (Fig. 5). Treatment of K562 cells with Fe (III)NTA also resulted in a slight increase in ferrochelatase activity. These results indicate that the expression of ferrochelatase may be positively regulated by iron.
Ferrochelatase activity in K562 cells decreased when these erythroleukemia cells were treated with iron chelators, desferrioxamine and Tiron, where intracellular iron levels decreased. In addition, expression of ferrochelatase was increased by the addition of Fe (III)NTA. These results signify that ferrochelatase activity is under the control of intracellular iron and possibly correlates with formation of the Fe/S cluster at the C-terminal region. The role of the Fe/S cluster was confirmed by observations that the activity of E. coli ferrochelatase, which lacks a Fe/S cluster, was insensitive to desferrioxamine when expressed in Cos7 cells. A previous study  revealed that recombinant ferrochelatase, as well as purified bovine ferrochelatase, showed an absorption spectrum similar to the Fe/S cluster. Disruption of the cluster with NO led to a loss of enzyme activity [7,11]. A mutated ferrochelatase F417S, a position at the C-terminal region, from a patient with erythropoietic protoporphyria exhibited only 2% of the activity of controls . Furthermore, ferrochelatase truncated at the C-terminal region lost enzyme activity completely [7,11]. Because the apoprotein of ferrochelatase is susceptible to proteolytic degradation (Fig. 2), the Fe/S cluster in mammalian ferrochelatase is a prerequisite for maintenance of the enzyme structure as well as activity. Thus, our present observations first demonstrate that the function of the Fe/S cluster of mammalian ferrochelatase is coupled with iron metabolism.
The stability of ferrochelatase purified from rat and bovine liver was different to that of bacteria. Namely, purified mammalian ferrochelatase stabilizes in the presence of dithiothreitol and glycerol . In contrast, E. coli ferrochelatase is very stable only in the presence of 2.5 mm EDTA which might perhaps protect it from proteolytic degradation . We are now able to explain the difference in stability between mammalian and bacterial enzymes, it is because the mammalian enzyme contains a labile Fe/S cluster. The Fe/S cluster of the mammalian ferrochelatase may have a unique physiological role in addition to catalytic activity. In this study, we demonstrated an additional function for this enzyme in that there is iron-dependent regulation of enzyme expression in mammals.
Yeast mitochondrial intermediate peptidase (MIP) functions in maturation of ferrochelatase and other iron-utilizing proteins, and is activated by iron in vitro[23,24]. A defect in yeast MIP leads not only to a loss of proteolytic processing of mitochondrial proteins, such as ferrochelatase and subunit IV of cytochrome oxidase, but also to mitochondrial iron depletion [24,25]. Our findings that the mature form of ferrochelatase in pCD-HF-transfected Cos7 cells was decreased markedly by treatment with desferrioxamine, suggested that MIP homologs in Cos7 cells may be nonfunctional under the iron-depleted conditions. We suspected that treatment of cells with desferrioxamine results in an accumulation of the precursor or intermediate form of ferrochelatase, but the present data showed that the amount of the precursor form in desferrioxamine-treated cells was similar to that in control cells (Fig. 4C). Thus, it is difficult to clarify whether transport of the apoprotein into mitochondria is suppressed by iron-depletion or how to promote degradation of ferrochelatase under iron-impaired conditions. Further investigations are required to determine whether mammalian mitochondrial processing protease or its homolog is involved in iron-dependent regulation of mitochondrial protein maturation.
The import of iron into mitochondria must be tightly regulated because little iron accumulation is observed in mice defective in heme biosynthesis , but the mechanism of iron metabolism in mitochondria is unclear. Recently, three yeast proteins involved in iron homeostasis within mitochondria have been described: the yeast ABC transporter Atm1p of the inner membrane , the matrix protein frataxin (Yfh1p) [24,27] and Ssq1p . Defects in any of these proteins cause large increases in mitochondrial iron content, and these proteins may contribute to the export of iron from the mitochondira. Furthermore, defective yeast Atm1p decreases cytosolic Fe/S cluster proteins, but not mitochondrial Fe/S cluster proteins . In contrast, in patients with Friedrich's ataxia, in which there is a defect in frataxin, levels of mitochodrial Fe/S cluster enzymes, such as aconitase and respiratory complexes I, II and III, were low compared with controls . When we compared the effect of iron depletion on the expression of ferrochelatase located in mitochondria and cytosol, the decrease in the enzyme located in the mitochondria by iron depletion was more sensitive than that in cytosol, thereby indicating that the level of the mitochondrial protein is liable to be influenced by intracellular iron levels. These results suggest that iron strongly regulates mitochondrial function via formation of the Fe/S cluster and heme.
This study showed that the heme content of cells decreased slightly when K562 cells were treated with desferrioxamine, indicating that the decrease in ferrochelatase causes a decrease in heme synthesis. However, an accumulation of protoporphyrin does not occur in desferrioxamine-treated K562 cells in which heme formation has been suppressed (Fig. 3). Moreover, no accumulation of any other porphyrins was observed (S. Taketani & Y. Nakahashi, unpublished results). In contrast, defects in ferrochelatase cause the accumulation of protoporphyrin in patients with erythropoietic protoporphyria [2,3,22], which is not the case in our findings. These results suggest that iron-depletion suppresses porphyrin biosynthesis in these human erythroleukemia cells, and this is consistent with the previous observation  that in erythroid cells, protoporphyrin IX levels are coupled to iron availability as a result of translational induction of erythroid specific ALA-S (ALA-S2) by iron. Because ALA-S2 mRNA contains an iron-responsive element at its 5′-UTR, translation of ALA-S2 mRNA depends on the availability of iron . In addition, Kim et al.  reported that NO-induced heme loss occurred in a rat primary culture of hepatocytes, in which ALA-S activity, as well as ferrochelatase, were decreased markedly. Intracellular iron is a major target of NO , and metabolism of the mitochondrial iron is impaired by NO , indicating that hepatic porphyrin biosynthesis can also be regulated by iron. Based on these observations, the availavbility of iron in the mitochondria regulates expression of the two enzymes, ferrochelatase and ALA-S, resulting in the control of heme biosynthesis in erythroid and nonerythroid cells.
This study was supported in part by grants from the Ministry of Education, Science, Sports and Culture of Japan. We thank Dr K. Miyamoto for the kind gift of pFC3, and K. Yasaka for the excellent technical assistance.