Overexpression of mitochondrial ferritin causes cytosolic iron depletion and changes cellular iron homeostasis

Ferritins are ubiquitous, highly conserved iron storage proteins that play a critical role in cellular and organismal iron homeostasis. In mammalian cells, cytosolic ferritin consists of 2 subunits, the H and L chains. Twenty-four subunits assemble to form a spherical shell that can contain up to 4500 atoms of iron in a mineral and compact form.^1,2  In ferritin shells, the H subunit has ferroxidase activity that converts soluble ferrous ions into inert ferric hydroxides. The L subunit lacks ferroxidase activity but is more efficient than the H subunit in inducing iron nucleation and mineralization within the shells.^3,4  Ferritin H and L chain mRNAs have identical iron-responsive elements (IREs) at their 5′-untranslated regions (UTRs) that bind to cytosolic iron regulatory proteins IRP1 and IRP2 by an iron-dependent mechanism.^2,5  When the level of iron in the so-called “labile iron pool” (LIP) is low, translation of ferritin mRNA is suppressed by the binding of IRPs to IREs, whereas iron supplementation up-regulates ferritin synthesis through the conversion of IRP1 to cytosolic aconitase and the degradation of IRP2.^5-8 

Mitochondria have a key role in iron metabolism because heme and various iron-sulfur (Fe-S) cluster-containing proteins are synthesized in them.^9,10  The last step in heme biosynthesis, the insertion of Fe²⁺ into protoporphyrin IX by ferrochelatase, occurs in the mitochondrial matrix.⁹  Fe-S clusters are synthesized mainly, if not entirely, in mitochondria and are combined with mitochondrial apo-proteins to form mature proteins or are exported from mitochondria for use by cytosolic and nuclear proteins.¹⁰  Mitochondrial iron level must be well regulated because an inadequate supply of iron would impair the metabolic and respiratory activities of the organelle, whereas excess “free” iron in mitochondria would promote the generation of harmful reactive oxygen species, which are produced as a side reaction of mitochondrial electron transport.^11,12 

Defects in mitochondrial iron transport and utilization can result in mitochondrial iron overload. There is extensive iron accumulation in erythroblast mitochondria of patients with X-linked sideroblastic anemia due to defective erythroid-specific 5-aminolevulinic acid synthase (eALAS)^9,13-15  and those with ringed sideroblasts associated with myelodysplastic syndrome.^13,15  Mitochondrial iron overload has also been documented in patients with Friedreich ataxia with defective frataxin^16-18  and in those with sideroblastic anemia with ataxia from defects in the Fe-S transporter ABC7.¹⁹  In addition, studies with yeast, the best-studied eukaryotic model of Fe-S cluster synthesis, showed that defects in any of the enzymes of the Fe-S cluster assembly pathway caused mitochondrial iron accumulation and lack of normal mitochondrial function.²⁰ 

The recent discovery of mitochondrial ferritin (MtFt) has contributed to our understanding of mitochondrial iron metabolism^21-24  and sideroblastic anemia. Although most cells, including normal erythroid cells, express extremely low levels of MtFt, it is highly expressed in erythroblast mitochondria (“ringed sideroblasts”) of patients with sideroblastic anemia^21,24  and it was shown that the iron in ringed sideroblasts is in MtFt.²⁴  In humans, MtFt is encoded by an intronless gene on chromosome 5q23.1. The protein is expressed as a 30-kDa precursor that is targeted to mitochondria. After intramitochondrial deletion of the N-terminal leader sequence, the 22-kDa MtFt subunits, which are highly homologous to H subunits, assemble into ferritin shells with ferroxidase activity.^21,22  In HeLa cells, overexpression of human MtFt, which incorporates iron, results in decreased cytosolic ferritin and increased transferrin receptor levels, suggesting cytosolic iron deficiency.²² 

Although the regulation of MtFt expression and its functions are not understood, it is likely that it plays a role in cellular iron homeostasis. To investigate the role of MtFt in cellular iron uptake, distribution, and mobilization, we generated a human cell line with tetracycline (tet)–controlled expression of mouse MtFt with which we provide direct evidence that MtFt dramatically affects intracellular iron metabolism.

EXPRE³⁵S³⁵S protein-labeling mix and ⁵⁹FeCl₃ were from Perkin Elmer (Boston, MA). All other reagents were purchased from Sigma (St Louis, MO) unless otherwise specified. The iron chelator salicylaldehyde isonicotinoyl hydrazone (SIH) was synthesized as described by Ponka et al.²⁵ 

The mouse MtFt cDNA encoding the full opening reading frame was amplified by polymerase chain reaction (PCR) using 5′-AGGGAATTCACCATGGGCCTGTCCTGCTTTTGGTTCTTCTC-3′ and 5′-GGCGGATCCTATTTAAGCGTAATCTGGAACATCGTATGGGTAGTGCTTGCTCTCGCTTCCAA-3′ primers with mouse testis QUICK-Clone cDNA library (Clontech, Mississauga, ON, Canada) and cloned into the pGEM-T vector (Promega, Madison, WI). This produced a plasmid containing a 768-bp fragment containing the entire mouse MtFt sequence, a C-terminal hemagglutinin (HA) epitope and an EcoRI site at the 5′ end and a BamHI site downstream of the stop codon at the 3′ end of MtFt. The 768-bp EcoRI/BamHI inserts were excised and subcloned into the EcoRI/BamHI site of pUHD10-3 (from Dr K. Pantopoulos, Lady Davis Institute, Montreal, QC, Canada) that contains a tetracycline-inducible hCMV minimal promoter²⁶  and pcDNA3.1(-) (Invitrogen, Burlington, ON, Canada) to yield pUHD-MtFt and pcDNA3.1(-)–MtFt, respectively. Constructs with the correct sequences were confirmed by sequencing.

The pUHD-MtFt plasmid was cotransfected with the puromycin-resistant pBabe plasmid into tTA-H1299 cells²⁷  by Lipofectamine Plus (Invitrogen). Stable transfectants were selected in medium containing 2 μg/mL puromycin, 250 μg/mL G418, and 2 μg/mL tetracycline. Representative clones were isolated for further characterization. The cells were maintained in Dulbecco modified Eagle medium supplemented with 10% (vol/vol) tetracycline-free fetal bovine serum (Clontech), 2 mM glutamine, and 100 U/mL penicillin, and 0.1 ng/mL streptomycin. MtFt expression was induced by removal of tetracycline (tet-off system).²⁷ 

The MtFt leader sequence was predicted using MitoProt II software²⁸  and the report by Corsi et al.²²  A truncated MtFt sequence (deletion of the first encoded 56 amino acid residues, Δ56-MtFt) was PCR-amplified by the primer pair of 5′-GGCGGATCCTATTTAAGCGTAATCTGGAACATCGTATGGGTAGTGCTTGCTCTCGCTTCCAA-3′ and 5′-AGGGAATTCACCATGGGCGACTCCACTAGACCTTCCAG-3′. Then the 603-bp fragment was subcloned into pcDNA3.1(-) and pUHD10-3 as described (see “Plasmid constructs”) for the full-length MtFt sequence. Transient transfection with pcDNA3.1(-)–MtFt and pcDNA3.1(-)–Δ56-MtFt were performed with the Lipofectamine Plus reagent. Protein expression was examined using Western blot analysis.

Cells were lysed with RIPA buffer (50 mM Tris [tris(hydroxymethyl)aminomethane]–HCl, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate [SDS]) or Munro buffer (10 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid], pH 7.5, 3 mM MgCl₂, 40 mM NaCl, 5% glycerol, 1 mM dithiothreitol, and 0.2% Nonidet P-40) and the lysates were immediately boiled in Laemmli loading buffer for 10 minutes. Protein content was determined by the Bradford assay (Bio-Rad, Hercules, CA). Equal amounts of protein were loaded, resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to a nitrocellulose blotting membrane (Pall, Pensacola, FL). The blots were blocked by incubating them for 1 hour with 5% nonfat milk in phosphate-buffered saline (PBS) containing 0.1% Tween 20 (PBST) and hybridized with anti-HA rabbit polyclonal IgG (Y11; Santa Cruz Biotechnology, Santa Cruz, CA), antitransferrin receptor (anti-TfR; Zymed, South San Francisco, CA), antiferritin (Dako, Carpinteria, CA), or anti–β-actin (Sigma, Oakville, ON, Canada). After washing 3 times for 15 minutes each with PBST, the blots were incubated for 1 hour with horseradish peroxidase-linked secondary antibody. Dilutions were 1:1000 for primary antibodies and 1:10 000 for antirabbit or antimouse IgG secondary antibodies. Peroxidase-coupled secondary antibodies were detected with Western lighting (Amersham, Buckinghamshire, United Kingdom).

A gel retardation assay, as described previously,²⁹  was used to measure the interaction between IRPs and IREs. Briefly, 6 × 10⁶ cells were lysed at 4°C in 100 μL Munro buffer and the lysates were centrifuged for 10 minutes at 10 000g to remove nuclei and unlysed cells. Samples of cytoplasmic extracts were diluted with Munro buffer without 0.2% Nonidet P-40 to a protein concentration of 1 μg/μL, and 10-μg aliquots were incubated with an excess amount of ³²P-labeled pST18-fer RNA transcript (containing one IRE).³⁰  This RNA was transcribed in vitro from a linearized plasmid template using T7 RNA polymerase in the presence of α-³²P] UTP. To form RNA-protein complexes, cytoplasmic extracts were incubated for 10 minutes at room temperature with this probe. Heparin (5 μg) was added and the incubation was continued for another 10 minutes to prevent nonspecific binding. RNA-protein complexes were analyzed in 6% nondenaturing polyacrylamide gels. In parallel experiments, samples were treated with 2% β-mercaptoethanol (β-ME) before adding the RNA probe to maximally activate IRP1-binding activity.

Human apo-transferrin (Tf) was labeled with ⁵⁹Fe as described previously.³¹  Equal numbers of cells (3 × 10⁶ cells) were seeded without or with tetracycline, cultured for 48 hours, and then incubated with 1 μM ⁵⁹Fe-Tf at 37°C for 24 hours. Cells were washed 3 times with ice-cold PBS, and radioactivity in the cell pellet and culture medium was measured in a γ counter (Packard Cobra II auto-gamma; Perkin-Elmer, Wellesley, MA).

After labeling with ⁵⁹Fe-Tf at 37°C for the indicated times, the cells were harvested and lysed at 4°C with lysis buffer (0.14 M NaCl, 0.1 M HEPES, pH 7.4, and 1.5% Triton X-100), and the lysates were analyzed by electrophoresis followed by autoradiography as described previously.^32,33  Briefly, following centrifugation at 10 000g for 10 minutes at 4°C, equal amounts of lysate samples were loaded onto native gradient gels (3%-20% polyacrylamide). After 4 hours of electrophoresis, the gels were dried and exposed to x-ray film.

Cells were incubated for 1 hour with 100 μCi/mL (3.7 MBq/mL) ³⁵S-methionine and ³⁵S-cysteine and lysed with RIPA buffer. Lysates containing equal amounts of protein were first treated with 30 μL protein A-Sepharose (Amersham, Piscataway, NJ) to remove proteins that bind nonspecifically to the beads. Then, 5 μg antiferritin antibody (Roche, Penzberg, Germany) was added to the supernatants, which were then incubated for 3 hours at 4°C. Protein A-Sepharose suspension (60 μL; 50% beads in RIPA buffer) was added to immunoprecipitate the ferritin-IgG complexes. The beads were washed 3 times with cold RIPA buffer (10 minutes each) and then boiled for 5 minutes with Laemmli loading buffer. Immunoprecipitated proteins were separated by 12.5% SDS-PAGE. The gels were treated with Amplify (Amersham), dried, and analyzed by autoradiography with a Phosphor Imaging Screen and a Storm (840) Scanner (Molecular Dynamics, San Diego, CA).

Both cytosolic and mitochondrial aconitase activities were measured by isocitric dehydrogenase-coupled nicotinamide adenine dinucleotide phosphate (NADPH) formation as previously described.³⁴  Briefly, 100 μL extracts containing equal amounts of protein was added to 800 μL of a solution containing 330 mM Tris-HCl, pH 7.4, 1.5 mM manganese sulfate, 0.6 mM NADP, and 1 U/mL isocitrate dehydrogenase. The reaction was started by adding 100 μL of 20 mM sodium citrate, and production of NADPH was monitored by measuring the increase in absorbance at 340 nm. One unit of aconitase is defined as the amount that converts 1 μmol citrate per hour to isocitrate at pH 7.4, at 25°C.

Mitochondria were isolated by an established method.³⁵  Briefly, 1 × 10⁸ cells were homogenized in a glass-Teflon homogenizer containing isolation medium (32 mM sucrose, 1 mM EDTA [ethylenediaminetetraacetic acid], and 10 mM Tris-HCl, pH 7.4). The homogenate was centrifuged at 1500g for 10 minutes to remove nuclei and unlysed cells. The supernatant was centrifuged at 17 000g for 10 minutes at 4°C and the pelleted mitochondria were washed twice with the isolation buffer. The postmitochondrial fraction was used as cytosolic fraction. Cross-contamination of cytosol and mitochondria was determined on the basis of the amount of activity of marker enzymes for cytosol (lactate dehydrogenase [LDH], EC 1.1.1.27)³⁶  and mitochondria (succinate dehydrogenase [SDH], EC 1.3.99.1).³⁷  Mitochondria contamination of cytosol was less than 0.5% (percent LDH activity in mitochondria/total cell homogenate) and cytosolic contamination of mitochondria was less than 1% (percent SDH activity in postmitochondrial fraction/total cell homogenate in RIPA buffer). Total nonheme iron was measured using the ferrozine method described by Fish³⁸  in cells grown for 72 hours without or with tetracycline. Cellular heme and nonheme ⁵⁹Fe were measured by an acid precipitation method.³⁹  Briefly, uninduced and induced clone B9 cells were grown for 2 days, ⁵⁹Fe-Tf was added, and the cells were cultured for an additional day. The cells were washed 3 times in PBS and lysed by adding 0.2 M HCl, and the lysates were boiled for 10 minutes to dissociate nonheme iron from proteins. ⁵⁹Fe-heme was precipitated by adding trichloroacetic acid to a final concentration of 7% and incubated at 4°C for 15 minutes. After centrifugation at 2000g for 5 minutes, the pellets were washed 3 times with 7% trichloroacetic acid and radioactivity in the pellets (heme ⁵⁹Fe) and supernatants (nonheme ⁵⁹Fe) was measured in a γ counter.

The sequence of the cloned mouse MtFt was aligned with the published sequence of mouse MtFt (NM_026286) and 2 variations were found. At nucleotide 37 there was an A>G substitution in the mitochondrial leader sequence and at nucleotide 488 there was a G>A substitution in the ferroxidase center (numbered from ATG start codon). The first resulted in a Ser → Gly change and the second to a Gly → Glu change (G107E; Glu is well conserved in human H, L, and mitochondrial ferritins, and many animals, plants, and bacteria ferritins^1,21 ). MitoProt II software²⁸  predicted that, although it would be preferred in the mitochondrial target sequence, the Ser replacement by Gly would not apparently decrease mitochondrial import efficiency.

Human H1299 cells expressing the transactivator tTA were transfected with cDNAs encoding HA-tagged MtFt under the control of a tetracycline-inducible promoter (tet-off system).²⁷  MtFt protein expression was analyzed by Western blotting using anti-HA tag antibody. Stable clones were screened and representative clones were selected (Figure 1A). Clone B9 was used for further characterization. Removal of tetracycline from the medium for 12 to 96 hours resulted in a profound induction of MtFt expression (Figure 1B). Expression of MtFt protein was time dependent and reached a maximum after 72 hours of induction. In agreement, a previous report about human MtFt,²²  mouse MtFt from stable transfectants migrated as a single band with a molecular mass of approximately 22 kDa on SDS-PAGE. On the other hand, from transient transfection with pcDNA 3.1-MtFt, 2 bands, the precursor and the mature protein, were detected by Western blotting. The precursor was about 28 kDa and was processed to the 22-kDa mature protein (Figure 1C). Transient transfection of cDNA with Δ56-MtFt resulted in nearly the same molecular size band corresponding to mature MtFt (Figure 1C). This is consistent with removal of the predicted leader sequence eliminated at the predicted position. Figure 2 illustrates the distribution of MtFt in clone B9 cells. Most of the MtFt, if not all, localized to the mitochondrial fraction (M1 and M2), but not in the postmitochondrial fraction (C1 and C2). The MtFt was concentrated about 10 times in the mitochondrial fraction compared with whole-cell homogenates (T1 and T2). This result suggests that the precursor of MtFt is targeted to mitochondria where it is processed to the mature protein.

IRPs sense the level of intracellular iron in the LIP and coordinate the expression of ferritin and Tf receptor (TfR) through binding to IREs located in the 5′-UTR of ferritin mRNA and the 3′-UTR of TfR mRNA.^2,5-8  We examined the effect of MtFt expression on IRP/IRE-binding activity. Cells were grown in the presence or absence of tetracycline for 24 to 72 hours, and lysates were then analyzed by gel retardation assays using an IRE probe. Removal of tetracycline led to a significant increase in IRE-binding activity in cells expressing MtFt, about 2.8- and 1.8-fold (quantified by densitometry) after 48 or 72 hours of induction, respectively (Figure 3A, lanes 1-6). As expected, there was no change in total IRP activity that was revealed when cell extracts were treated with β-ME (Figure 3A, lanes 7-12). Because in human tissues and cells, both IRP1/IRE and IRP2/IRE complexes comigrate in gel retardation assays, here they are referred to as “IRP.”

Next, we examined the effect of overexpression of MtFt on the expression of TfR. Clone B9 cells were grown in the presence or absence of tetracycline for 24 to 72 hours and TfR levels were analyzed by Western blotting. The increased expression of MtFt correlated with a time- and dose-dependent increase in TfR levels (2.3- and 2.8-fold after 2 and 3 days of induction, respectively; Figure 3B), whereas the levels of β-actin did not change. This increase in the levels of TfR after expression of MtFt is consistent with a cellular response to iron deficiency. To explore this, we examined the role of MtFt in the expression of cytosolic ferritin, which is regulated at the posttranscriptional (translational) level by IRE-IRP interactions.^5-9,40  After cells were labeled metabolically with ³⁵S-methionine and ³⁵S-cysteine, ferritin synthesis was analyzed by immunoprecipitation (Figure 3C). Expression of MtFt after the removal of tetracycline clearly inhibited ferritin synthesis (Figure 3C, lanes 2, 4, and 6) even after 24 hours of induction. We also measured the steady-state level of ferritins (Figure 3D) in the cells with the same treatment. A decrease in cytosolic ferritin levels was found in the cells with 24-hour induction of MtFt expression; after 48-hour induction of MtFt, the cytosolic ferritin was barely detectable. It appears that MtFt promoted an apparent iron-deficient phenotype in H1299 cells, namely, increased IRP-binding activity, increased TfR levels, and inhibited cytosolic ferritin synthesis.

We next examined the role of MtFt expression in the uptake of ⁵⁹Fe from ⁵⁹Fe-Tf and the distribution of ⁵⁹Fe between cytosol and mitochondria. Cells grown for 0, 24, or 48 hours with or without tetracycline were incubated with ⁵⁹Fe-Tf for another 24 hours and cell-associated radioactivity was measured. Expression of MtFt on removal of tetracycline resulted in a significant (P < .01) increase in ⁵⁹Fe uptake from ⁵⁹Fe-Tf by approximately 180% and 250% at 48 hours and 72 hours of induction, respectively (Figure 4A). Analysis of the distribution of radioactivity between cytosol and mitochondria showed that MtFt-expressing cells had a significantly higher percentage of the ⁵⁹Fe taken up in mitochondria. At 48 hours of induction, it increased from 17.5% in uninduced cells to 33.0% in cells expressing MtFt. After 72 hours of induction, the percentage of ⁵⁹Fe in mitochondria was about 27% in cells expressing MtFt, still significantly higher than in the control cells (16.4%; Figure 4B). Most, if not all, of this increased ⁵⁹Fe was likely in MtFt. In another experiment, it was found that MtFt increased total cellular nonheme iron from 0.264 ± 0.009 nM/10⁶ cells to 0.358 ± 0.012 nM/10⁶ cells after 72 hours of induction of MtFt expression. Thus, MtFt expression led to cytosolic iron deficiency and increased total iron uptake with more iron shunted into the mitochondria.

To further examine the effect of MtFt on cellular iron distribution, we used native PAGE combined with ⁵⁹Fe autoradiography to evaluate the distribution of ⁵⁹Fe in the major cellular iron-containing molecules. Figure 5A shows that MtFt (confirmed by supershift assay; data not shown) migrated more slowly than cytosolic ferritin but their positions were not far apart, suggesting that MtFt probably forms multimeric shells similar to cytosolic ferritin. Induction of MtFt expression almost totally blocked iron from being incorporated into cytosolic ferritin, most of it being found in MtFt, whereas in uninduced cells the majority of the ⁵⁹Fe was taken up by cytosolic ferritin (Figure 5A). These results suggest that MtFt forms functional ferritin-like shells and competes with cytosolic ferritin for iron. There also was an increase in ⁵⁹Fe-Tf-TfR complexes after 48 and 72 hours of induction of MtFt expression, whereas the amount of ⁵⁹Fe in the low-molecular-weight fraction, referred to as fraction “Y” in a previous study,³²  did not apparently change (Figure 5A). The positions of the Tf and Tf-TfR bands was verified by supershifts using anti-Tf antibodies (data not shown). This, together with the low incorporation of ⁵⁹Fe into cytosolic ferritin, suggests that most of the iron from Tf shunted into mitochondria may bypass the cytosol. Next, the cells were incubated with ⁵⁹Fe-Tf for 24 hours, washed, and cultured for 48 hours without ⁵⁹Fe and without tetracycline to induce MtFt expression. Figure 5B clearly shows that the induction of MtFt leads to the mobilization of iron from both cytosolic ferritin and fraction Y, indicating that cytosolic iron can also be incorporated into MtFt.

SIH, which mobilizes iron from reticulocytes whose mitochondria are loaded with nonheme iron,²⁵  was used to examine whether iron in MtFt can be chelated (Figure 5C-D). Uninduced and induced (48 hours) cells were incubated with ⁵⁹Fe-Tf for 24 hours and with SIH for another 5 hours, after which ⁵⁹Fe in the medium and cells was measured. Figure 5C shows that SIH released 33% of the total cellular ⁵⁹Fe from uninduced cells, whereas only 18% of ⁵⁹Fe was released from MtFt-expressing cells. Similarly, SIH released about 31% (quantified by densitometry) of ⁵⁹Fe from cytosolic ferritin, whereas only 21% was released from MtFt in induced cells (Figure 5D).

Because IRP1 and cytosolic aconitase are the same protein with 2 different functions and because mitochondrial aconitase is an Fe-S cluster protein whose mRNA contains a functional IRE in its 5′-UTR,^41,42  we examined whether MtFt-induced cytosolic iron depletion directly affects the activities of these 2 Fe-S cluster proteins. MtFt expression was induced for 72 hours and isocitric dehydrogenase-coupled NADPH formation was used to measure aconitase activities of both proteins. Consistent with the previous result showing increased IRP-binding activity on MtFt induction (Figure 3A), cytosolic aconitase activity decreased about 50% (P < .01) in induced cells (Figure 6A). Interestingly, mitochondrial aconitase also responded to MtFt induction and was decreased by about 30% (P < .05) in induced cells (Figure 6B).

In another experiment, uninduced and induced clone B9 cells were grown for 2 days, ⁵⁹Fe-Tf was added, and the cells were cultured for an additional day. After washing, ⁵⁹Fe was measured in the heme and nonheme iron fractions. The ratios of heme-⁵⁹Fe to nonheme ⁵⁹Fe were 0.110 ± 0.013 and 0.086 ± 0.003 (P = .03) in uninduced and induced cells, respectively, suggesting that in cells expressing MtFt less iron is available for heme synthesis.

Iron is a precious metal for organisms and is involved in a broad spectrum of essential biologic functions. However, some aspects of iron chemistry diminish its appeal. Unless bound to specific ligands, at pH 7.4 and physiologic oxygen tension, iron would exist only in the forms of virtually insoluble ferric hydroxides. Moreover, the catalytic action of iron in one-electron redox reactions can play an important role in the formation of toxic free radicals that ultimately cause oxidative damage to vital cell structures. Therefore, organisms are equipped with specific proteins designed for this metal's acquisition, transport, and storage in a soluble nontoxic form.

Levi et al²¹  recently identified a new intronless gene that encodes an H-ferritin–like protein, MtFt, that is expressed only in mitochondria. Most cells and tissues, including normal erythroblasts, express very low levels of MtFt.²¹  However, MtFt levels increase dramatically in ringed sideroblasts from patients with sideroblastic anemia.^21,24  It is well known that a ringed sideroblast is a pathologic erythroid precursor containing excessive deposits of nonheme iron with characteristic perinuclear distribution accounting for the ring appearance.^13-15  However, neither the pathogenesis of mitochondrial iron accumulation in ringed sideroblasts nor the function of MtFt is known.

Previous work by Corsi et al²²  has documented that the overexpression of human MtFt in HeLa cells leads to a decrease in cytosolic ferritin levels and enhances TfR expression. Our results confirm and expand on their work. In this study we cloned the mouse MtFt gene, which was then transfected into a cell line under the control of a tetracycline-responsive promoter.²⁷  We showed that removal of tetracycline in clone B9 cells caused a time-dependent induction of MtFt, starting as early as 12 hours and reaching maximum at 72 hours (Figure 1B). MtFt expressed in stable transfectants migrated as a single band with a molecular mass of about 22 kDa, corresponding to the mature mitochondrial protein. However, transient transfectants expressed not only mature MtFt but also a 28-kDa protein, likely a precursor of MtFt expressed in mitochondria (Figure 1C). As expected, transient transfection with cDNA lacking mitochondrial leader sequence generated protein with a molecular size corresponding to that of mature MtFt (Figure 1C).

Exploiting the “tet-off” system in stable transfectants, we confirmed that the expression of MtFt leads to an increase in TfR levels (Figure 3B) and a total suppression of cytosolic ferritin synthesis (Figure 3C-D) that confirms a decrease in cytosolic ferritin levels described earlier.²²  However, in addition to this we have also demonstrated, for the first time, that the overexpression of MtFt causes a dramatic increase in IRP binding to IREs (Figure 3A), a finding that provides direct evidence of cytosolic iron deficiency. In another set of experiments we documented that MtFt-expressing cells, as compared with their wild-type counterparts, took up Tf-borne ⁵⁹Fe with significantly higher rates (Figure 4A) which is likely the result of increased TfR expression. Moreover, mitochondria from induced cells contained about twice as much radioactive iron at 48 hours (Figure 4B) and their total nonheme iron content increased by approximately 36% at 72 hours compared with controls. In contrast to uninduced cells in which the majority of the radioactive iron was in cytosolic ferritin, induced cells had virtually no ⁵⁹Fe in cytosolic ferritin and contained virtually all their radioactive iron in MtFt (Figure 5A). Moreover, the induction of MtFt very effectively mobilized ⁵⁹Fe from the cytosolic ferritin of uninduced cells that had previously been incubated with ⁵⁹Fe-Tf; in fact, most cellular ⁵⁹Fe following the induction ended up in MtFt (Figure 5B). These experimental results demonstrate not only a remarkable capacity of MtFt to acquire iron from cytosolic sources, but also a very efficient translocation of Tf-derived iron into MtFt. Although the mechanism underlying this is unknown, the MtFt-inducible system described here could be exploited to characterize, at least in part, a yet unknown pathway involved in mitochondrial iron use. In addition, the rapid mobilization of iron from cytosolic ferritin to MtFt in our inducible system will also be a useful tool in investigating the nature of the release of iron from ferritins in situ, a mechanism that has been one of iron metabolism's enduring dilemmas.

It was previously shown that a 1-hour incubation of reticulocytes with ⁵⁹Fe-Tf in the presence of porphyrin synthesis inhibitors leads to a dramatic accumulation of nonheme ⁵⁹Fe in mitochondria^39,43  that is available for chelation by SIH; this chelator, at 200 μM, removed approximately 45% of ⁵⁹Fe per hour from reticulocytes with mitochondrial nonheme ⁵⁹Fe accumulation.⁴⁴  It has also been shown that certain 2-pyridylcarboxaldehyde isonicotinoyl hydrazone analogues can effectively mobilize this nonheme iron accumulated in reticulocyte mitochondria.⁴⁵  The chemical form in which ⁵⁹Fe accumulates in reticulocyte mitochondria is unknown, but the radioiron is unlikely inserted in MtFt because this protein is not expressed in normal erythroid cells. The efficacy with which SIH mobilizes ⁵⁹Fe from reticulocytes suggests that mitochondrial membranes do not represent much of a barrier for this chelator. We deemed it important to examine the capacity of SIH to mobilize ⁵⁹Fe from cells that contain most of their radioactive iron either in cytosolic or mitochondrial ferritin. We found that SIH mobilized ⁵⁹Fe less efficiently from cells that contained ⁵⁹Fe in MtFt than from those having ⁵⁹Fe stored in cytosolic ferritin (Figure 5C). Similarly, SIH removed ⁵⁹Fe much more readily from cytosolic than mitochondrial ferritin (Figure 5D). This decreased availability to iron chelation was also observed in HeLa cells overexpressing human MtFt, when using desferrioxamine to mobilize the iron stored in MtFt.²² 

We also investigated whether or not the expression of MtFt affects some proteins in which iron is essential for their function. Our data clearly demonstrate that the induction of MtFt significantly inhibited the activity of both cytosolic and mitochondrial aconitase (Figure 6), enzymes that require an intact [4Fe-4S] cluster for their function. It needs to be pointed out that a lower activity of mitochondrial aconitase activity may, at least in part, be due to decreased synthesis of this enzyme because its mRNA contains an IRE in its 5′-UTR.^41,42  Moreover, our data suggest that in MtFt-expressing cells iron is less available for heme synthesis than in wild-type cells.

In summary, this study has demonstrated that MtFt is capable of very effectively acquiring endosomal, Tf-borne iron as well as cytosolic iron that is inserted in ferritin. Moreover, the induction of MtFt compromises the function of at least some nonheme iron proteins and also appears to interfere with iron delivery to heme. Furthermore, compared with cytosolic ferritin, iron inserted into MtFt is less available for chelation by the strong chelator SIH. Collectively, our data suggest that the avidity of MtFt for, and its association with, iron is stronger than those of its cytosolic counterpart. These properties of MtFt could be potentially useful in diminishing oxidative damage in conditions associated with mitochondrial iron overload, such as Friedreich ataxia^16,17  or X-linked sideroblastic anemia with ataxia.¹⁹ 

We thank Drs Kostas Pantopoulos and Jian Wang for the tTA-H1299 cells and Dr Lukas Kuhn for pST18-fer plasmid.