Cells Lacking Pfh1, a Fission Yeast Homolog of Mammalian Frataxin Protein, Display Constitutive Activation of the Iron Starvation Response*♦

Background: Defects in the protein frataxin give rise to Friedreich ataxia. Results: A new Friedreich ataxia model using fission yeast has been generated and its phenotype and proteome characterized. Conclusion: Frataxin absence triggers a complete iron starvation program, sufficient to generate all the associated respiratory defects. Significance: Our new model system may contribute to decipher the role of frataxin. Friedreich ataxia is a genetic disease caused by deficiencies in frataxin. This protein has homologs not only in higher eukaryotes but also in bacteria, fungi, and plants. The function of this protein is still controversial. We have identified a frataxin homolog in fission yeast, and we have analyzed whether its depletion leads to any of the phenotypes observed in other organisms. Cells deleted in pfh1 are sensitive to growth under aerobic conditions, display increased levels of total iron, hallmarks of oxidative stress such as protein carbonylation, decreased aconitase activity, and lower levels of oxygen consumption compared with wild-type cells. This mitochondrial protein seems to be important for iron and/or reactive oxygen species homeostasis. We have analyzed the proteome of cells devoid of Pfh1, and we determined that gene products up- and down-regulated upon iron depletion in wild-type cells are constitutively misregulated in this mutant. Because of the particular signaling pathway components governing the iron starvation response in fission yeast, our experiments suggest that cells lacking Pfh1 display a decrease of cytosolic available iron that triggers activation of Grx4, the common regulator of the iron starvation gene expression program. Our Schizosaccharomyces pombe Δpfh1 strain constitutes a new and useful model system to study Friedreich ataxia.

Friedreich ataxia is an inherited autosomal recessive disease causing degeneration in the central and peripheral nervous system, cardiomyopathy, skeletal abnormalities, and increased risk of diabetes mellitus (1)(2)(3)(4). A decreased expression of a highly conserved nucleus-encoded mitochondrial protein, known as frataxin, causes Friedreich ataxia. In 98% of the cases, an unstable hyper-expansion of a GAA triplet repeat in the first intron of the gene is the most common genetic mutation (2,5). The expanded GAA repeat impairs frataxin transcription by adopting an abnormal triple helical structure.
This disease is thought to be the consequence of a mitochondrial defect related to iron metabolism. Thus, the cardiac tissues from patients with Friedreich ataxia contain iron deposits, are deficient in respiratory complexes I-III and aconitase activities, and have reduced mitochondrial DNA (6,7). The finding of iron accumulation in the hearts of both patients and mouse models (6,8,9) suggests that excess free iron is involved in the production of reactive oxygen species and is responsible for the oxidative damage to iron-sulfur clusters (ISCs) 3 and loss of mitochondrial DNA (10).
Iron is a vital metal for most biological organisms and participates in an astonishing array of biological reactions such us DNA synthesis, cell cycle progression, and respiration (11). Nevertheless, this redox-active transition metal presents a dilemma to cells, because iron can also catalyze the deleterious oxidation of biomolecules via Haber-Weiss/Fenton chemistry when combined with reactive oxygen species (ROS) (12). Accordingly, the concentration of iron in cells is tightly regulated by control of its uptake and intracellular storage (13).
Frataxin is a mitochondrial iron-binding protein, and its primary function remains controversial, because iron homeostasis, intracellular fluctuations of ROS, oxidation and damage of ISCs, and mitochondrial dysfunctions are all intimately linked, and it is not trivial to establish the sequential order of events leading to the disease. Homologs to human frataxin have been found in bacteria, fungi, and plants (14). In particular, unicellular model systems are being exploited to determine the origin of the pleiotropic phenotypes displayed by patients. In particular, Saccharomyces cerevisiae cells deficient in the frataxin homolog YFH1 are unable to carry out oxidative phosphorylation, lose mitochondrial DNA (15), display impaired iron efflux out of mitochondria with a consequent accumulation of iron in this compartment (16), suffer iron depletion in the cytosol, show elevated expression of high affinity iron uptake, exhibit heme deficiency (17), and have an increased sensitivity to oxidative stress. Additional properties of these ⌬YFH1 mutant cells include defects in the synthesis of ISCs with a consequent deficiency of ISC-containing proteins and loss of respiratory competence. Furthermore, in vitro studies indicate that YFH1 can bind iron and deliver it to Isu1 (ISC assembly scaffold protein) (18,19). The phenotypes caused by decreased levels of frataxin point to the notion that it plays a role in ISC synthesis and mitochondrial and cellular iron misregulation with a consequent oxidative stress. However, studies on mice and yeast also show that the loss of ISCs precedes iron accumulation (9,20), so that the anomalies may not be a direct result of iron-induced oxidative damage. Furthermore, studies using an S. cerevisiae model of conditional expression of frataxin demonstrated that the primary consequence of frataxin depletion is to trigger upregulation of the iron transport system before affecting ironsulfur enzyme activities (21).
The success of previous simple models of Friedreich ataxia prompted us to investigate whether we could isolate a Schizosaccharomyces pombe strain mimicking the phenotypes observed in other organisms. We have found an open reading frame (ORF), pfh1, coding for the fission yeast homolog of human frataxin. We have constructed a strain carrying a deletion of the gene, and we found phenotypes resembling those of other Friedreich ataxia model systems. Despite the great amount of data obtained with previous model systems, the analysis of the proteome of our frataxin-deficient strain has provided us with new data and has prompted us to discard previous hypotheses, such as the possible role of frataxin on ISC biogenesis.

EXPERIMENTAL PROCEDURES
Alignment of Frataxin Sequence-The alignment of the frataxin protein sequence of S. pombe, S. cerevisiae, and human isoform 1 was performed by multiple sequence alignment with hierarchical clustering software (22).
Solid Sensitivity Assay-To analyze sensitivity to different agents on plates, S. pombe strains were grown and spotted as described (30). Serially diluted cells were spotted into minimal medium or YE5S plates containing or not the indicated concentrations of hydrogen peroxide (H 2 O 2 ), deferoxamine mesylate (Sigma), or iron (FeCl 3 ⅐6H 2 O, Sigma). The spots were allowed to dry, and the plates were incubated at 30°C during 2-3 days under aerobic or anaerobic conditions.
Labeling of Total Disulfides for One-dimensional Electrophoresis Analysis-Protein extracts of exponentially growing S. pombe cells were obtained and labeled as described previously (31).
Colorimetric Assay for Iron Quantification-Yeast cells were grown in YE5S anaerobic liquid cultures (50 ml). Cells were washed twice with PBS buffer, pH 7.4, and treated as described (32) using ferrozine (Fluka). Absorbance of iron-chelator complex was recorded at A 565 in a UV-visible Ultraspec 2100-pro (Amersham Biosciences) spectrophotometer. The accuracy of the assay was improved by subtracting nonspecific absorbance recorded at A 680 . The number of cells was calculated from A 600 (A 600 0.5 ϭ 1 ϫ 10 7 cells/ml). Standard curves were prepared from 10 to 40 nmol of FeCl 3 dissolved in 3% nitric acid. All chemicals (except ferrozine from Fluka) were purchased from Sigma and resuspended in ultrapure water obtained from a Millipore Milli-Q Advantage. Data were obtained from three independent experiments and are expressed as mean Ϯ S.E.
Protein Carbonylation-The detection method was performed as described (33).
Enzymatic Activity Assay for Aconitase-To perform aconitase enzymatic activity assay for wild-type and ⌬pfh1 strains, S. pombe cells were grown in YE5S in anaerobic conditions to an A 600 of 0.5. Aconitase activity was assayed as described (34) following method 2. Aconitase activity can be measured spectrophotometrically at 340 nm using citrate as the substrate of aconitase, and the isocitrate formed was then converted to ␣-ketoglutarate by NADP ϩ -dependent isocitrate dehydrogenase. Yeast cells were washed once with PBS buffer, pH 7.4. Total protein extracts were prepared by homogenization with glass beads in Tris buffer (50 mM Tris-HCl, pH 7.6, 1 mM cysteine, 1 mM citrate, 0.5 mM MnCl 2 ). Insoluble material was removed by centrifugation during 10 min at 16,000 ϫ g at 4°C. Supernatants were collected, and 10 l of the total extract were mixed with 90 l of reaction buffer (50 mM Tris-HCl, pH 7.4, 30 mM sodium citrate, 0.5 mM MnCl 2 , 0.2 mM NADP ϩ , isocitrate dehydrogenase (2 units/ml, Fluka)). Absorbance at 340 nm (⑀ 340 ϭ 6.22 mM Ϫ1 cm Ϫ1 ) was recorded by UV-visible spectrophotometer (UV-1700 Pharma Spec Shimadzu). Data were obtained from three independent experiments and are expressed as mean Ϯ S.E.
Measurement of Oxygen Consumption-Oxygen consumption of ϳ10 7 cells in 1 ml, collected from anaerobic cultures at an A 600 of 0.5, was measured as described before (35).
Fluorescence Microscopy-Cells expressing Pfh1-GFP were grown in YE5S until an A 600 of 0.5. They were then incubated with 0.1 g/ml MitoTracker Red CMXRos (Invitrogen) during 30 min. Cells were pelleted and resuspended in YE5S. Fluorescence microscopy and image capture was performed as described previously (36).
Cell Extracts and Immunoblot Analysis of Pap1-Preparation of S. pombe trichloroacetic acid (TCA) protein extracts to measure Pap1 concentration was performed as described before (38). Samples were separated by 8% SDS-PAGE. Gels were transferred to membranes, and they were probed with polyclonal anti-Pap1 antiserum and anti-Sty1 antiserum (as loading control) (36) Growth Curves-Yeast cells were grown in YE5S, and the cultures of NG60, NG147, and NG148 strains were grown anaerobically. The initial A 600 of the growth curves was 0.1, and recording of the growth curves was performed as described (30).
Quantification of Proteins by Dimethyl Labeling-50 ml of wild-type and ⌬pfh1 cells were grown to exponential phase (A 600 1.0) anaerobically. Cells were passed to aerobic conditions with shaking for 3 h, which did not significantly affect viability of the cultures (supplemental Fig. S1). Cells were collected after addition to the cultures of 100% TCA to a final concentration of 10%; pellets were washed with 20% TCA, and cells were lysed by vortexing with glass beads in 250 l of 12.5% TCA. Cell lysates were then pelleted, washed twice in cold acetone, and dried. Each pellet was resuspended in 500 l of 200 mM Tris-HCl, pH 8.5, 6 M urea, 5 mM EDTA, and 0.05% SDS. Protein concentration was determined by Bradford assay. 25 g of protein of each sample were reduced by addition of 15 l of 10 mM DTT in 200 mM triethylammonium bicarbonate (1 h at 37°C) and then alkylated by adding 15 l of 20 mM iodoacetamide in 200 mM triethylammonium bicarbonate (30 min at room temperature). Upon reduction and alkylation, protein extracts were diluted in 200 mM triethylammonium bicarbonate to a final urea concentration of 1 M. At this point, protein extracts were trypsinized (Promega), and tryptic peptides were labeled with either normal (wild type) or deuterium-labeled (⌬pfh1) formaldehyde (Sigma), respectively (39,40). A modification was added to the dimethyl labeling protocol (40); the reaction mixture was not acidified prior to mixing, but instead 8 l of 1% ammonia solution was added, as described (31). All the details regarding mass spectrometers used and data processing have been published elsewhere (31) Analysis of Dimethyl-labeled Proteins-For the analysis of dimethyl-labeled proteins, Proteome Discoverer version 1.2.0.208 (Thermo Fisher Scientific, Bremen, Germany) was used to extract MS/MS spectra that were queried using Mascot version 2.3, as described (31). Protein ratios are reported as the median of the measured peptide ratios for a given protein.
Regarding data analysis, database searches of MS/MS spectra were extracted as reported previously (31). For protein accession and gene identification (supplemental Tables S2-S5), peptides were searched against S. pombe GeneDB (Wellcome Trust Sanger Institute).

RESULTS
Identification and Deletion of the Gene pfh1, Which Codes for the Fission Yeast Frataxin Homolog-We searched the S. pombe genome for genes with homology to the human frataxin gene, and found an ORF (SPCC1183.03c) coding for a 158-amino acid-long polypeptide sharing 42 and 44% of protein identity with human and S. cerevisiae frataxin, respectively (Fig. 1A). We named the protein Pfh1 (pombe frataxin homolog 1), following the nomenclature used for the S. cerevisiae frataxin (yeast frataxin homolog, YFH1). YFH1 diverged from the human gene earlier in evolution than S. pombe pfh1 (Fig. 1A), which further motivated us to analyze the effect of pfh1 deletion on the cell's physiology. We generated a strain lacking the whole ORF by genetic recombination, as shown by Northern blot analysis (Fig.  1B), and we had to perform the selection of the deleted clones under semi-anaerobic conditions. The substitution of the whole pfh1 ORF by an antibiotic resistance cassette was checked by PCR (data not shown). When anaerobic cultures of strain ⌬pfh1 were spotted on agar plates, we observed severe growth defects in the presence of oxygen (Fig. 1C). In fact, even the viability of liquid cultures of strain ⌬pfh1 was severely compromised when grown under aerobic conditions, as shown in Fig. 1D. We also determined that cells lacking Pfh1 display sensitivity to exogenous oxidative stress (Fig. 1E). These results are consistent with an essential role of Pfh1 in cellular fitness, because it is required for respiratory growth and for survival in front of intrinsic ( Fig. 1, C and D) or extrinsic (Fig. 1E) accumulation of ROS.
Phenotypic Characterization of Strain ⌬pfh1-Because our ⌬pfh1 strain seemed to share the sensitivity to aerobic growth of other Friedreich ataxia model systems, we tested whether it also recapitulated some of the characteristic features related to iron homeostasis and mitochondrial metabolism. We detected over 3-fold increases in total iron in extracts from strain ⌬pfh1 compared with wild-type cells ( Fig. 2A). The strain, either before or after a transient shift to aerobic conditions, displayed hallmarks of oxidative stress, such as increased levels of reversibly oxidized thiols (Fig. 2B) or protein carbonylation (Fig. 2C). The activity of the ISC-containing protein aconitase was also significantly reduced in the mutant (Fig. 2D), and respiratory competence was also severely affected as determined by measuring oxygen consumption (Fig. 2E). All these features have also been reported for previous models of Friedreich ataxia and confirmed the relevance of our newly developed system. Consistently, protein Pfh1 displayed mitochondrial localization, as reported previously for human and S. cerevisiae frataxin proteins (Fig. 2F).
Characterization of the Proteome of ⌬pfh1 Cells-To decipher the role of Pfh1 in the cell's physiology, we analyzed the protein composition of total cell extracts of cells lacking Pfh1 and of wild-type cells, when anaerobic cultures were shifted for 3 h to aerobic conditions (supplemental Fig. S1). Tryptic peptides derived from cell extracts were labeled with either normal (wild type) or deuterium-labeled (⌬pfh1) formaldehyde. Peptides were then mixed and analyzed by LC-MS/MS, as described under "Experimental Procedures." The complete list of proteins misregulated in ⌬pfh1 cells is provided in supplemental Tables S2 and S3. A total of 58 proteins were overexpressed more than 1.5-fold, and 71 were down-regulated more 2-fold, in extracts from ⌬pfh1 cells when compared with wildtype extracts ( Table 1). The expression of more than half of these misregulated proteins (39 with lower expression and 35 with higher expression; Table 1) has not been identified as regulated by a common transcription factor, as far as we know. However, the expression of 52 proteins (32 with lower expression and 20 with higher expression; Table 1) is regulated upon iron starvation, at least at the gene level (41). Furthermore, nine of the proteins overexpressed in ⌬pfh1 cells are regulated by the transcription factor Pap1 (Table 1), at least at their mRNA levels (42).
Cells Lacking Pfh1 Display Increased Expression of Some Pap1-dependent Proteins-At least nine proteins whose expression depends on the transcription factor Pap1 are over-represented in extracts from cells lacking Pfh1 (supplemental Table S4). In response to H 2 O 2 , the transcription factor Pap1 up-regulates transcription of genes required for adaptation to oxidative stress and for tolerance to toxic drugs. H 2 O 2 induces oxidation of Pap1, its nuclear accumulation, and expression of more than 50 Pap1-dependent genes, some of which are antioxidant genes and others drug resistance genes (Fig. 3A) (36,42). We have recently reported that the ability of Pap1 to bind and activate the second subset of promoters, those of drug tolerance genes, is independent of Pap1 oxidation. Thus, nuclear localization of nonoxidized Pap1 or overexpression of the protein is sufficient to trigger activation of drug resistance genes, but Pap1 has to be not only nuclear but also oxidized to bind to another transcription factor, Prr1, and activate antioxidant genes (43).
We first checked by Northern blot analysis whether the expression of the genes coding for some of the Pap1-dependent proteins overexpressed in ⌬pfh1 extracts are also up-regulated under basal conditions. As shown in Fig. 3B, the levels of Pap1dependent drug resistance mRNAs are up-regulated by aerobic growth even prior to the addition of H 2 O 2 in cells lacking Pfh1, but not the antioxidant ones such as trr1. Because the presence of Pap1 at the nucleus can also be accomplished by overexpression of the transcription factor, with the subsequent activation of drug tolerance genes (43), we tested whether the mRNA for pap1 (Fig. 3C) and/or the levels of Pap1 protein (Fig. 3D) were elevated in strain ⌬pfh1. Surprisingly, both pap1 mRNA (Fig.  3C) and Pap1 protein levels (Fig. 3D) are 3-10-fold higher in pombe and S. cerevisiae with human isoform1 frataxin sequence. Red letters represent the consensus sequence of amino acids that are conserved in the three organisms, and blue letters indicate the amino acid sequence that is present just in two of the organisms. The phylogenetic tree relative to the frataxin gene for these organisms is represented in the right panel. B, Northern blot analysis of pfh1 mRNA. Total RNA from strains 972 (WT) and NG60 (⌬pfh1) was obtained from cultures growing anaerobically. The RNA was analyzed by Northern blot using a pfh1 probe. Total ribosomal RNA was used as a loading control. C, cells lacking Pfh1 displays severely compromised aerobic growth. Strains 972 (WT) and NG60 (⌬pfh1) were grown anaerobically in YE5S to a final A 600 of 0.5, and serial dilutions from 10 5 to 10 cells were spotted in duplicate into YE5S plates and incubated at 30°C for 2-3 days in aerobic or anaerobic conditions. D, viability of wild-type (WT) and NG60 (⌬pfh1) cultures in response to aerobic growth. Strains 972 (WT) and NG60 (⌬pfh1) were grown overnight (ON) in YE5S in aerobic (ϩ ON ϩ O 2 ) or anaerobic (ϩ ON Ϫ O 2 ) conditions, and then serial dilutions of logarithmic phase cells were spotted as described in C; plates were grown under anaerobic conditions (ϪO 2 plate). E, strain ⌬pfh1 is sensitive to H 2 O 2 stress. Survival of wild-type (WT), AV18 (⌬sty1), and NG60 (⌬pfh1) strains in response to the indicated concentration of H 2 O 2 in plates under anaerobic conditions is shown. DECEMBER 14, 2012 • VOLUME 287 • NUMBER 51

JOURNAL OF BIOLOGICAL CHEMISTRY 43045
this strain background than in wild-type cells, which could explain the constitutive activation of some drug resistance genes (Fig. 3E). Further work will be required to understand how up-regulation of pap1 levels is occurring upon deletion of pfh1; this is the first reported genetic mutation that affects Pap1 activity at the level of transcription. Oxidation of the transcription factor is not altered in this mutant (data not shown), which discards enhanced H 2 O 2 levels triggering its activation. Table S4, the genes of 20 proteins overexpressed in strain ⌬pfh1 are normally triggered in wild-type cells upon iron starvation, at least at their mRNA levels (41). In S. pombe, addition of chelators such as dipyridyl (DIP) to cell cultures induces pronounced changes in the gene expression program, which are dependent on Grx4, Fep1 and Php4 (Fig. 4A). Briefly, the inactivation of Fep1 and activation of Php4 transcriptional repressors mediate the cellular response to iron deficiency, by either up-or down-regulating, respectively, the expression of genes (44). Thus, upon low iron conditions Fep1 is released from promoters of genes involved in iron uptake, such as fio1 or str3 (Fig. 4, A and B) (45). As shown in Fig. 4B, the expression of the Fep1 repressor-dependent fio1 and str3 genes is de-repressed under basal conditions in cells lacking Pfh1, similarly to what occurs in cells devoid of the Fep1 repressor. Consistently, a double deletion ⌬pfh1 ⌬fep1 strain has the same phenotype as cells lacking only Pfh1. It is important to point out that ⌬fep1 cells are not sensitive to grow under aerobic conditions, as shown on solid plates (Fig. 4C) or in liquid cultures (Fig. 4D), which indicates that the constitutive activation of iron uptake genes observed in ⌬pfh1 cells does not contribute to their sensitivity to grow under aerobic conditions. Cells Lacking Pfh1 Display Decreased Levels of Proteins Normally Down-regulated upon Iron Starvation-As shown in supplemental Table S5, 32 proteins down-represented in extracts from strain ⌬pfh1 are also down-regulated at their gene levels upon iron starvation, most of them in a Php4-dependent man- Oxidized thiol labeling of wild-type (WT) and NG60 (⌬pfh1) extracts was analyzed by fluorescent one-dimensional gel electrophoresis (see "Experimental Procedures") in anaerobic or aerobic conditions. Silver staining was used as control of protein loading. C, protein carbonylation levels are high in extracts from ⌬pfh1 strain. Strains 972 (WT) and NG60 (⌬pfh1) were grown aerobically for 3 h to a final A 600 of 0.5, and protein carbonylation was detected as described in "Experimental Procedures." Silver staining was used as a control of protein loading. D, activity of the ISC-containing protein aconitase is impaired in a ⌬pfh1 strain. Aconitase activity of 972 (WT) and NG60 (⌬pfh1) strains was performed as indicated in "Experimental Procedures." E, respiratory rate of strain ⌬pfh1 is diminished. Oxygen consumption of 972 (WT) and NG60 (⌬pfh1) cells grown in YE5S under anaerobic conditions was measured as indicated in "Experimental Procedures." Data in A, D, and E were obtained from three independent experiments and are expressed as mean Ϯ S.E. Significant differences in A, D, and E between wild-type and ⌬pfh1 cells were determined by Student's t test, (*, p Ͻ 0.05, and **, p Ͻ 0.01). F, mitochondrial localization of Pfh1-GFP protein. Strain NG142 (expressing Pfh1-GFP) was grown under aerobic conditions. Cells were incubated with the mitochondrial dye Mitotracker Red prior to analysis. The cellular distribution of the fusion protein (middle panels; Pfh1-GFP) and the dye (top panels; Mitotracker) was determined by fluorescence microscopy. The same cells under differential interference contrast (Nomarski) optics are shown in the bottom panels. a Number of proteins down-regulated is more than 2-fold in ⌬pfh1 versus wildtype cells after 3 h under aerobic conditions. b Number of proteins up-regulated is more than 1.5-fold in ⌬pfh1 versus wild-type cells after 3 h under aerobic conditions. c The genes coding for these proteins are up-regulated upon iron starvation (41), probably in a Fep1-dependent manner (52). d Six of these proteins are included in the set of proteins induced by Pap1. e The genes coding for these proteins are down-regulated upon iron starvation, 23 of them in a Php4-dependent manner (41). f The genes coding for these proteins are up-regulated in response to nontoxic doses of H 2 O 2 in a Pap1-dependent manner (42). g Number of proteins for which regulation of expression is unknown. FIGURE 3. Cells lacking pfh1 overexpress a subset of Pap1-dependent genes. A, scheme of the activation of Pap1 pathway in S. pombe wild-type cells. Pap1 activates two subsets of genes, the antioxidant and the drug resistance genes. In wild-type cells, oxidation of Pap1 upon H 2 O 2 stress induces its nuclear (NUC) accumulation, and a heterodimer with Prr1 is formed, which is able to activate both sets of promoters, the antioxidant (trr1, srx1, and ctt1) and the drug resistance (obr1, caf5, and c663.08c) genes. CYT, cytosol. B, Northern blot analysis of Pap1-dependent genes. Total RNA from strains 972 (WT) and NG60 (⌬pfh1) was obtained from cultures growing anaerobically or shifted to aerobic conditions for the times indicated and treated or not with H 2 O 2 . The RNA was analyzed by Northern blot using probes for the Pap1-dependent genes zwf1, trx1, tpx1, obr1, caf5, 663.08c, and trr1. Total ribosomal RNA was used as a loading control. C, Northern blot analysis of pap1 gene. Total RNA from strains 972 (WT) and NG60 (⌬pfh1) strains was obtained from cultures growing anaerobically or aerobically. The pap1 ORF was used as a probe. D, amount of Pap1 protein is higher in a ⌬pfh1 than in wild-type cells. Total TCA protein extracts were analyzed by Western blot with anti-Pap1 antibodies; anti-Sty1 antibodies were used as a loading control. E, hypothetical scheme of Pap1 activation in ⌬pfh1 cells. There is an accumulation of reduced Pap1 that is able to enter in the nucleus and activate the drug resistance genes. Grx4 is a glutaredoxin whose role is to sense the iron levels in cells and transmit the iron status to two repressor factors, Fep1 and Php4. Fep1 is the repressor of the iron uptake genes, and Php4 is a protein that represses a set of genes that code for iron-containing proteins and iron storage. B, Northern blot analysis of Fep1-dependent genes. Total RNA from strains 972 (WT), NG60 (⌬pfh1), NG1 (⌬fep1), and NG147 (⌬pfh1 ⌬fep1) was obtained from cultures growing in YE5S aerobically treated or not with 250 M of the iron chelator DIP for 90 min. C, aerobic growth defects of strain ⌬pfh1 are not due to the increment in the iron uptake system. Strains 972 (WT), NG60 (⌬pfh1), NG1 (⌬fep1), and NG147 (⌬pfh1 ⌬fep1) were grown in YE plates under aerobic or anaerobic conditions, as described in Fig. 1C. D, growth curves of 972 (WT), NG60 (⌬pfh1), and NG1 (⌬fep1) strains were done in YE5S liquid medium, and A 600 was recorded at the indicated times for each culture during 35 h.

Cells Lacking Pfh1 Display Increased Expression of Proteins Normally Up-regulated upon Iron Starvation-As shown in supplemental
ner. Thus, upon low iron, Php4 accumulates at the nucleus and represses transcription of genes coding for iron storage proteins or iron-consuming proteins (Fig. 4A) (46). First of all, we confirmed by Northern blot that not only protein levels but also mRNA levels for some Php4-dependent genes, such as pcl1 and isa1, are constitutively low in ⌬pfh1 cells (Fig. 5A). It is important to point out that many of the genes down-regulated by Php4 upon iron starvation are ISC-or iron-containing proteins, and many of them are essential (supplemental Table S5). Because deletion of the php4 gene, coding for a repressor, leads to the lack of iron starvation-dependent gene repression (Fig.  5A) (41), we tested whether the oxygen-sensitive phenotype of cells lacking Pfh1 could be alleviated by further deletion of the php4 gene. As shown in Fig. 5A, expression of pcl1 and isa1 was constitutively up-regulated in the double knock-out strain, as expected. Furthermore, the sensitivity to grow under aerobic conditions was partially suppressed in the double mutant as shown both on solid plates (Fig. 5B) and in liquid cultures (Fig.  5C). Excess iron or addition or iron chelators did not significantly improve or impair the growth of ⌬pfh1 cells (Fig. 5B). Therefore, the constitutive down-regulation observed in strain ⌬pfh1 of some Php4-dependent genes, many of which are essential, is partially, but not totally, contributing to the severe phenotype of our Friedreich ataxia model. It is also worth mentioning that genes coding for the ISC-containing proteins aconitase are also Php4-dependent and are therefore down-regulated in our ⌬pfh1 strain; this fact would be sufficient to justify the low activity levels of this and other ISC-containing proteins in this genetic background (Fig. 2D).
Transcriptome of ⌬pfh1 Cells Resembles That of ⌬grx4 ⌬fep1 Cells-As explained above, both the transcriptome (as observed by Northern blots) and the proteome of cells lacking Pfh1 are very similar to that of wild-type cells under iron limitation. Thus, iron import is exacerbated, and iron storage and iron-containing proteins are down-regulated. Because the glutaredoxin Grx4 is the common link between up-regulation (Fep1-dependent) and down-regulation (Php4-dependent) of transcription under iron deprivation (Fig. 4A), our frataxin homolog could be participating in signaling by regulating Grx4. However, deletion of the gene coding for this glutaredoxin does not lead to the same transcriptome changes as deletion of pfh1 (Fig. 6A). Indeed, the only strain that misregulates the iron regulon in the same directions as strain ⌬pfh1 is the double deletion ⌬grx4 ⌬fep1 (Fig. 6A). We interpret these results by hypothesizing that the absence of Pfh1 does not directly participate in the signaling cascade leading to the complete iron starvation gene program; instead, it may really trigger an iron starvation situation, which should consequently activate Grx4. It is worth mentioning that ⌬grx4 ⌬fep1 cells also display severe growth phenotypes in the presence of oxygen (Fig. 6B). That indicates that many of the phenotypes of our Friedreich ataxia model system, strain ⌬pfh1, are caused by the misregulation of the iron regulon itself, and in particular by the constitutive Php4-dependent repression of many essential ISC-containing proteins.

DISCUSSION
Deficiencies in frataxin give origin to Friedreich ataxia. We have here developed a new model system to study the molecular events leading to the disease; fission yeast cells lacking Pfh1  display all hallmarks of other previously reported model systems. Importantly enough, the function of frataxin (and therefore the molecular events leading to the disease in the absence of this protein) is still controversial, and our new model system may shed light onto the essential function of this protein under aerobic conditions. S. cerevisiae YFH1, the budding yeast frataxin homolog, was earlier reported to directly participate in ISC biogenesis, because aconitase activity was reduced in cells lacking the protein, and this inactivation seemed to precede iron accumulation (9,47). In particular, yeast frataxin has been suggested to participate in ISC maturation as an iron donor, based on its reported interaction with Isu1 (19,48). However, it has also been demonstrated that restricting oxidative damage, either by decreasing ROS production (49) or by diminishing available iron (50), prevents aconitase inactivation. Furthermore, a conditional knockdown of YFH1 expression has allowed establishing the order of sequential events occurring upon frataxin depletion, indicating that induction of iron import is a primary event leading to the late inactivation of ISC-containing proteins (21). Our results suggest that the low levels of aconitase activity in cells lacking Pfh1 are a consequence of the earlier activation of the Php4 repressor, which in wild-type cells lowers expression of most ISC-containing proteins in an iron starvation-dependent manner. Therefore, our work does not support the idea that frataxin is required for ISC biogenesis, but rather directly or indirectly participates in iron sensing and signaling.
As explained above, the fission yeast gene expression program upon iron starvation is governed by Grx4, which ultimately is responsible of both up-and down-regulation of several genes. Downstream of Grx4, the inactivation of Fep1 and activation of Php4 transcriptional repressors mediate the cellular response to iron deficiency (44). Briefly, when iron is experimentally depleted by the use of chelators, Fep1 is released from promoters of genes involved in iron uptake (45), whereas Php4 accumulates at the nucleus and represses transcription of genes coding for iron storage or iron-consuming proteins (46). Importantly enough, Grx4 is the real sensor of iron deprivation, probably through its ISC. 4 However, the apoprotein does not mimic an iron starvation response, indicating that the loss of the ISC is not the mechanism by which wild-type Grx4 becomes active. 4 Similarly, cells devoid of Grx4 only mimic the iron starvation response with regard to gene down-regulation by Php4 but cannot trigger activation of Fep1-dependent genes (Fig.  6A). As we have shown here, cells devoid of Pfh1 display all hallmarks of an iron deprivation situation, which can hardly be accomplished by genetic modulation of iron sensing/signaling components. In fact, only cells carrying double deletion of grx4 and fep1 display a similar transcriptome and phenotype as ⌬pfh1 cells (Fig. 6, A and B). Pfh1 could modulate Grx4 activity by, for instance, stabilizing the inactive, iron-rich conformation via chaperone or scaffold properties, its deficiency leading to the basal accumulation of the iron starvation-induced conformation. However, it is difficult to reconcile this putative chaperone role of Pfh1 on Grx4 activity when the first protein has mitochondrial localization (Fig. 2F) and Grx4 displays cytoplasmic and nuclear localization. 4 Our results unambiguously indicate that the absence of S. pombe frataxin causes a real iron starvation situation able to trigger the complex up-and downregulation of the gene expression program. Indeed, activation of a complete iron starvation response, including up-regulation of the high affinity iron transport system Fet3-Ftr1 and increased rate of iron uptake, has been described before in the yeast model of Friedreich ataxia (10). Furthermore, both increased iron uptake and a decrease in the major pathways of mitochondrial iron utilization have been described in a mouse model of Friedreich ataxia (51). A possible role for frataxin, which nicely fits with these results, is its participation in the regulation of cellular iron homeostasis from the mitochondria. Maybe Pfh1 depletion triggers accumulation of the metal in this compartment, with the concomitant decrease of available cytosolic iron and Grx4 activation. Further experiments to confirm this hypothesis are of course required.
Our study also strongly suggests that constitutive repression of many essential ISC-containing proteins in strain ⌬pfh1 contributes to the severe phenotypes observed, because they can be partially suppressed by deletion of php4 (Fig. 5, B and C). Studies on our new S. pombe model system on Friedreich ataxia will hopefully contribute to understanding the function of frataxin and to easily test different therapeutic interventions, which may prevent the onset of the disease.