Evidence for the presence of ferritin in plant mitochondria Marco Zancani 1 , Carlo Peresson 1 , Antonino Biroccio 2 , Giorgio Federici 2 , Andrea Urbani 3 , Irene Murgia 4 , Carlo Soave 4 , Fulvio Micali 5 , Angelo Vianello 1 and Francesco Macrı ` 1 1 Dipartimento di Biologia ed Economia Agro-Industriale, Sezione di Biologia Vegetale, Universita ` di Udine, Italy; 2 Laboratorio di Biochimica Clinica, Ospedale Pediatrico del Bambino Gesu ` – IRCCS, Roma, Vatican State; 3 Centro Studi sull’Invecchiamento (Ce.S.I), Facolta ` di Medicina e Chirurgia, Dipartimento di Scienze Biomediche, Universita ` ‘G. D’Annunzio’, Chieti, Italy; 4 Dipartimento di Biologia, Sezione di Fisiologia e Biochimica delle Piante, Universita ` di Milano, Italy; 5 Dipartimento di Biochimica, Biofisica e Chimica delle Macromolecole, Universita ` di Trieste, Italy In this work, evidence for the presence of ferritins in plant mitochondria is supplied. Mitochondria were isolated from etiolated pea stems and Arabidopsis thaliana cell cultures. The proteins were separated by SDS/PAGE. A protein, with an apparent molecular mass of approximately 25–26 kDa (corresponding to that of ferritin), was cross-reacted with an antibody raised against pea seed ferritin. The mitochondrial ferritin from pea stems was also purified by immunopre- cipitation. The purified protein was analyzed by MALDI- TOF mass spectrometry and the results of both mass finger print and peptide fragmentation by post source decay assign the polypeptide sequence to the pea ferritin (P<0.05). The mitochondrial localization of ferritin was also confirmed by immunocytochemistry experiments on isolated mitochon- dria and cross-sections of pea stem cells. The possible role of ferritin in oxidative stress of plant mitochondria is discussed. Keywords: ferritin; iron; mitochondria; Arabidopsis thaliana; Pisum sativum. Iron is an essential element for all living organisms [1]. In green plants, its importance mainly derives from the presence at the active sites of metalloproteins involved in the electron transport chains linked to both oxygen evolution (photosynthesis) and consumption (respiration). However, iron(II) ions may also a mplify t he damaging effect of reactive oxygen species (ROS) o n membranes, proteins and nucleic acids [2]. This happens particularly during the response of plants to diseases and other environmental stresses accompanied by an excess of ROS production (oxidative stress) [ 3,4]. The intracellular concen- tration of free iron has therefore to be tightly controlled at both the uptake and storage levels [5]. In the plant cell, chloroplasts and mitochondria are two of the major sites of ROS generation [6,7]. In both cases, the direct transfer of one electron from the electron transport chain to oxygen (univalent reaction) generates superoxide anion, which then dismutates, spontaneously or enzymat- ically, to hydrogen peroxide. The latter can react with iron(II) ion (Fenton reaction) generating the highly reactive hydroxyl radical. To prevent this risk, plant cells have evolved two strategies, namely scavenging of hydrogen peroxide or sequestration of iron [2]. Chloroplasts possess both systems, the scavenging (e.g. ascorbate peroxidase) [6] and the iron-buffering proteins (ferritins) [8]. Conversely, plant mitochondria seem to have only systems to scavenge H 2 O 2 or to prevent its generation [7,9], but not to sequester iron. Only recently a mitochondrial ferritin, encoded by an intronless gene, has been described in erythroblasts of subjects with impaired h eme synthesis [10]. The gene, expressed in HeLa c ells, has permitted to reveal t hat exogenous iron is available to mitochondrial ferritin as it is to cytosolic ferritin [11]. Ferritins are highly conserved p roteins consisting of large multimeric shells that can store up to 4500 atoms of iron [12,13]. The latter is taken u p in the ferro us form and immobilized after oxidation c atalyzed by ferroxidase sites as ferric hydroxides or as amorphous hydrous ferric oxyphos- phate [12]. Iron c an also be released from the core of ferritins. This process is a ffected by r educing agents, but does not imply shell breakdown [13]. Nevertheless, an in vitro degradation of plant ferritin, induced by iron exchange, was described [14]. Therefore, ferritins can play a critical role in the cellular regulation o f iron storage and homeostasis. Soluble (ferritins) and insoluble (phytosiderin) iron- storage proteins have been described in dry pea seeds [14]. While animal ferritins are mainly cytosolic proteins, the plant ones appear to be localized in chloroplasts of plant cells or, more i n general, in plastids [8,15]. In this work it is shown t hat ferritins are also present in plant mitochondria. Materials and methods Isolation of Percoll-purified plant mitochondria Crude mitochondria (CMt) were isolated from etiolated pea (Pisum sativum L., cv. Alaska) stems as previously described [16], and purified by a P ercoll discontinuous gradient (PMt) Correspondence to F. Macrı ` , Dipartimento di Biologia ed Economia Agro-Industriale, Sezione di Biologia Vegetale, Universita ` di Udine, via Cotonificio 108, I-33100 Udine, Italy. Fax: +39 0432558784, Tel.: +39 0432558781/82, E-mail: biolveg@uniud.it Abbreviations: CMt, crude mitochondria; IDP, inosine 5¢-diphos- phate; MP, mitochiondrial matrix proteins; PAAF, polyclonal a nti- body against pea seed ferritin; PMt, Percoll-purified mitochondria; PSD, post source decay; ROS, reactive oxygen species; TOF, time-of-flight. (Received 24 June 2004, accepted 23 July 2004) Eur. J. Biochem. 271, 3657–3664 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04300.x as described in [17]. Where indicated, to obtain extremely pure pea mitochondria, PMt were subjected to a second discontinuous Percoll gradient. Matrix proteins (MP) were obtained from PMt as described in [16]. Mitoplasts (Mpl) were obtained from PMt after osmotic shock for 10 min in 10 m M HEPES/Tris (pH 7.6), 40 m M sucrose. Mitoplasts were collected from the pellet after centrifugation at 13 000 g for 10 min and resuspended in 20 m M HEPES/ Tris (pH 7.5), 0.25 M sucrose. Crude and purified mitochondria of Arabidopsis thaliana were isolated from liquid cell cultures resuspended in 10 mL of 10 m M HEPES/Tris (pH 7.6), 0.5 M mannitol, 10 m M EDTA, 2 m M cystein, homogenized with a Turrax at 4 °C, diluted with 20 mL of the above buffer, further homo- genized by a P otter homogenizer and centrifuged at 1000 g for 10 min. The supernatant was centrifuged at 13 000 g for 10 min and the final pellet was resuspended in 30 mL of 10 m M HEPES/Tris (pH 7.6), 0.5 M mannitol and centri- fuged at 2000 g for 5 min. The supernatant was collected and centrifuged at 13 000 g for 10 min. The final pellet (CMt) was resuspended in 20 m M HEPES/Tris, 0 .25 M sucrose. To obtain A. thaliana purified mitochondria, the final pellet was resuspended in 20 m M 3-[N-morpholino]pro- panesulfonic acid/KOH (pH 7.2), 0.3 M mannitol, 1 m M EDTA and handled as described for pea stem PMt [17]. Enzyme assay ATPase activities (vanadate-sensitive, marker enzyme for plasma membrane; molybdate-sensitive, marker enzyme for cytosolic soluble phosphatases; bafilomycin A 1 -sensi- tive, marker enzyme for tonoplast; oligomycin-sensitive, marker enzyme for mitochondria) were assayed as previ- ously described [18]. Latent IDPase (marker enzyme for Golgi), antimyc in A-insensitive cytochrome c reductase (marker enzyme for endoplasmic reticulum) and glucose- 6-phosphate dehydrogenase (marker enzyme for plastids) activities were detected as described in [19–21], respectively. Immunoprecipitation The immunoprecipitate was obtained from purified mito- chondria that had been frozen and thawed three times and then centrifuged at 12 000 g for 15 min. The super- natant ( 50 lL) was taken and 2 lL of rabbit polyclonal antibody raised against pea seed ferritin (PAAF, described in [22]) was added. After incubation for 1 h at 4 °C, 50 lL of protein-A sepharose (50% v/v slurry, washed twice in 50 m M Tris/HCl, pH 8.0), was added and incubated for 1 h at 4 °C. The immune complex was precipitated by centrifugation at 12 000 g for 20 s and the pellet washed thrice with 50 m M Tris/HCl (pH 8.0). The pellet was resuspended in 50 m M Tris/HCl (pH 7.5), 100 m M dithioerythritol and 1% (w/v) SDS, and then boiled at 95 °C for 3 min. The sample for electrophoresis analysis was obtained by collecting the supernatant ( 30 lL) after centrifugation at 12 000 g for 20 s and addition of 10 lLof75%(w/v)glycerolplus1lLof 0.1% (w/v) bromophenol blue. Analytical electrophoresis Gel electrophoresis was carried out in 12% (w/v) polyacryl- amide gels containing 0.1% (w/v) SDS [23]. After SDS/ PAGE, the gels were either stained with Coomassie Brilliant Blue R-250, or layered onto a nitrocellulose membrane to transfer the proteins by electroblotting. The nitrocellulose membranes were incubated with either PAAF or antibodies raised against the a/b-subunit of mitochondrial ATPase (1 : 5000 dilution) [24] and the reaction was developed by the activity of t he alkaline phosphatase conjugated to anti- (rabbit IgG) Ig. For the immunodecoration, in the presence of the monoclonal antibodies against cytochorme c (PharMingen International, 1 : 10 000 dilution), the reac- tion was developed by the activity of alkaline phosphatase conjugated to anti-(mouse IgG) Ig. The cross-reactivity with the antihuman mitochondrial ferritin (HuMtF) was performed as described in [11]. Table 1. Marker enzyme activity in crude (CMt) and purified pea mitochondria (PMt). The activity of antimycin A-insensitive cytochrome c reductase ( marker for endoplasmic reticulum) detected in pea microsomes, prepared as describe d in [40], was 570 nmolÆ(mg proteinÆmin) )1 ;the activity of the glucose-6-phosphate dehydrogenase (marker for plastids) detected in pea s tem etioplast, prepared as described in [ 41], was 235 nmol NADPH reduced (mg pro teinÆmin) )1 . n.d., Not determined. Marker enzyme CMt nmolÆ(mg proteinÆmin) )1 Percentage of control PMt nmolÆ(mg proteinÆmin) )1 Percentage of control ATPase 1m M ATP (control) 135 100 78 100 + 100 l M Na 3 VO 4 97 72 70 90 + 0.1 l M Bafilomycin A 1 125 93 77 98 + 100 l M Na 2 MO 4 146 108 93 120 +2lgÆmL )1 Oligomycin 70 52 20 26 Latent IDPase 1m M IDP 320 19 + 0.05% Brij 58 611 25 D 291 6 Cytochrome c reductase n.d. 457 +2l M Antimycin A n.d. 105 Glucose-6-phosphate dehydrogenase 35 6 3658 M. Zancani et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Mass spectrometry analysis Identification of polypeptides from polyacrylamide gel plugs was pursued by the trypsin mass fingerprint technique on a MALDI-TOF mass spectrometer. In short, the protein band was excised from a Coomassie-stained SDS/PAGE, cysteines were reduced and alkylated with iodoacetamide [25]. The samples were then digested with porcine trypsin (Promega) in 4 0 m M ammonium bicarbonate at 37 °Cfor 6–8 h. The reaction was sto pped by freezing the samples at )80 °C. Tryptic peptides were extracted by ZipTip C18 (Millipore) reverse phase material, d irectly e luted and crystallized in a 50% (v/v) acetonitrile/water saturated solution of a-cyano-4-hydroxycinnamic acid. MALDI mass spectra were recorded in the positive ion mode with delayed extraction on a Reflex IV time-of-flight instrument equipped w ith a multiprobe inlet and a 337 nm nitrogen laser. Mass spectra were obtained by averaging 50–200 individual laser shots. Calibration of the spectra was internally performed by a two-point linear fit using the autolysis products of trypsin at m/z ¼ 842.50 and m/z ¼ 2211.10. Database search with the peptide masses was performed against the NCBInr, taxon Viridiplantae, database using the peptide search algorithm MASCOT (Matrix Science). Fragments generated by post source decay (PSD) experi- ments were fitted using the database search algorithm MASCOT (Matrix Science) and analyzed by the de novo sequencing routine of B iotools (Bruker-Daltonik). Immunochemical electron microscopy Cross-sections of etiolated pea stem and isolated mitochon- dria were fixed with 4% (v/v) paraformaldehyde and 0.5% (v/v) glutaraldehyde in 0.17 M phosphate, 0.17 M sucrose buffer (pH 7.0) for 3 h at 4 °C. The cross-sections or the mitochondria were washed several times in 0.17 M phos- phate, then dehydrated in ethanol and embedded in LR White M acrylic resin (Sigma). Immunolabelling of ultra- thin sections (120 nm on 300 mesh nickel grids) was carried out by grids flotation technique at room temperature for 1 h on drops of blocking buffer: 1% (w/v) bovine serum albumin, 20% (v/v) normal goat serum in 0.1 M Tris- buffered saline (pH 7.4), and then incubated for 2 h in Tris- buffered saline (pH 7.4) containing PAAF (diluted 1 : 5), 1% (w/v) bovine serum albumin, 4% (v/v) fetal bovine serum, and 0.1% (v/v) Tween-20. After several washes in Tris-buffered saline to remove the antibody excess, the sections were incubated for 2 h in the same incubation medium, but at pH 8.4, containing secondary antibody gold-conjugated 10 nm goat anti-(rabbit IgG) Ig (British BioCell, Cardiff, UK) diluted 1 : 100. Finally, the sections were counterst ained with uranyl a cetate ( 2% w/v) for 3 min and l ea d citrate solution (0.25% w/ v) for 2 min and observed with Philips EM 208 electron microscope at 80 kV accelerating voltages. Anti-fe rritin Ig was omitted in t he controls. Fig. 1. Identification of ferritin in pea stem mitochond ria. (A) SDS/ PAGE(12%)analysisofproteins(25lg) from crude mitochondria (CMt), purified mitochondria (PMt), matrix from pea stem purified mitochondria (MP), and to tal pea seed proteins (CP , control p roteins); molecular mass of protein standards is indicated in kDa ( Std). (B) Immunoblotting of the same proteins with polyclonal antibod y against ferritin (PAAF). (C) Immunoblotting of recomb inant h uman mito- chondrial ferritin (rHuMt, 10 ng), protein extract from HeLa cells overexpressing human mitoc hondrial ferritin ( MtF-HeLa, 30 lg) and matrix proteins from pea stem purified mito chondria (MP, 35 lg) with antihuman mitochondrial ferritin polyclon al antibody after native 6% PAGE. (D) SDS/PAGE (12%) of the 25–26 kDa protein purified by immunoprecipitation. Table 2. Sequence coverage by trypsin digestion peptide mass fingerprint of the pea ferritin, purified from mitochondria, with t he translated sequence precursor of a pea ferritin (SwissProt accession number P19975). In bold are reported the protein regions covered in the mass fingerprint. Peptide sequences confirmed by fragmentation analysis by post source decay (PSD) are underlined. The putative peptide leader sequence located at the N-terminus is highlighted in black. 1 MALSSSKFSS FSGFSLSPVS GNGVQKPCFC DLRVGEKWGS RKFRVSATTA 51 PLTGVIFEPF EEVKKDYLAV PSVPLVSLAR QNFADECESV INEQINV EYN 101 ASYVYHSLFA YFDRDNVALK GFAKFFKESS EEHREHAEKL MKYQNTRGGR 151 VVLHPIKDVP SEFEHVEKGD ALYAMELALS LEKLTNEKLL NVHSVAERNN 201 DLEMTHFIEG EYLAEQVEAI KKISEYVAQL RRVGKGHGVW HFDQRLLHGV 251 HGA Ó FEBS 2004 Plant mitochondrial ferritin (Eur. J. Biochem. 271) 3659 Protein assay Protein concentration was determined by the method of Bradford [26], using bovine s erum albumin as a s tandard. Results Purified pea stem mitochondria, particularly when com- pared with CMt, were almost devoid o f contamination from different types of cellular components (Table 1). ATPase activity of this fraction was, indeed, uninhibited or only very slightly inhibited by vanadate (plasmalemma ATPase inhibitor), bafilomycin A 1 (tonoplast ATPase inhibitor) and molybdate (soluble phosphatase inhibitor), but strongly inhibited by oligomycin (mitochondrial ATPase inhibitor). In addition, PMt showed a low level of latent IDPase ( Golgi membrane marker enzyme). Cytochrome c reductase activ- ity was assayed in the presence and absence of antimycin A to assess the contamination from endoplasmic reticulum. In purified mitochondria (in the presence of antimy cin A), the activity was 4.35 times and almost six-fold lower than that recovered in c ontrol mitochondria and microsomes, respect- ively. On the other hand, the activity of cytochrome c reductase, still detected in the presence of antimycin A, could depend on the presence of a similar enzyme on the outer membrane of plant mitochondria [27]. Finally, this preparation exhibited a negligible glucose-6-phosphate dehydrogenase activity (plastid marker enzyme), partic- ularly when compared to that of a sample of etioplasts isolated from the same plant material. The proteins of CMt, PMt, and the relative matrix components were subjected to SDS/PAGE, in compari- son with a pea seed protein extract containing ferritin Fig. 3. Immunocytological l ocalization of ferritin in etiolated pea stem. (A) Cross-section of etiolated pea stem; cw, cell wall, v, vacuole, m, mitochondria. (B) and (C) Higher magnification of t he same electron micrograph showing labeled mitochondria. Arrows in dicate electron- dense particles after immunolabeling with PAAF followed by gold- conjugated se condary a ntibo dy. Bars co rres pond to 30 0 lm. Fig. 2. Localization of ferritin in pea stem purified mitochondria. (A) Immunoblotting with P AAF of PMt (25 lg) incubated (+) or not ( –) with 0.5% (w/v) Triton X-100 for 10 min, then subjected to proteolysis with 125 lgÆmL )1 trypsinfor30minat25°C and stopped by the addition of 1 m M PMSF. (B) Imm unoblotting of PMt (25 lg) and Mpl (25 lg) with monoclonal antibody raised against cytochrome c (Cyt c), polyclonal antibody raised a gainst the a/b-subunit of mito- chondrial ATPase (a/b-subunit) or PAAF (ferritin). 3660 M. Zancani et al. (Eur. J. Biochem. 271) Ó FEBS 2004 (Fig. 1A). Proteins, thus separated, were then subjected to an immunoblot assay by using PAAF (Fig. 1B). The results show that this antibody cross-reacted with a protein exhibiting an apparent molecular mass of approximately 25–26 kDa, a v alue similar to that o f ferritins. This reactivity w as achieved in all types of preparations. In particular, in some cases two very close bands were evident. As already suggested, they represent ferritin and a p roduct of degradation of the same protein [14]. The same cross- reactivity was a lso detected for a protein of twofold purified pea stem mitochondria (result not shown). In addition, the pea stem mitochondrial protein, after nondenaturing elct- rophoresis and blotting, cross-reacted with the polyclonal antibody anti-human mitochondrial ferritin [10] (Fig. 1C). The matrix proteins from purified pea mitochondria were also subjected to purification by immunoprecipitation. SDS/ PAGE analysis of the immunoprecipitated revealed, after Coomassie staining, a protein band at 25–26 kDa (Fig. 1D). The tryptic peptides of this band were analyzed with a MALDI-TOF m ass s pectrometer and the monoisotopic masses of each singly charge species were annotated with their intensities. These data were fitted on the NCBI non- redundant Viridiplantae database returning, with a confidence score greater than 95% (P<0.05) accuracy, the pea ferritin 1 chloroplast precursor (NCBI accession gi/417006; Swiss- Prot accession P19975). This assignment was confirmed by fragmentation analysis employing a MALDI-TOF post source decay experiment, selecting the ion species at 1078.52 amu (MH + ) with a time gate ion selector. The resulting fragmentation pattern was characteristic the y and b ion series of the sequence ISEYVAQLR (223–231). The PSD fragments were fitted on the NCBI nonredundant Viridi- plantae d atabase returning again the ferritin sequence ISEYVAQLR (223–231). The overall mass fingerprint data cover about the 30% of the assigned sequence and details are reported in Table 2. The theoric molecular mass, 23.6 kDa, calculated from the database sequence after removal of the N-terminus signal peptide, is in agreement with the value of 25–26 kDa estimated from the SDS/PAGE. The localization of ferritin in pea stem purified mito- chondria was investigated (Fig. 2). Figure 2A shows an immunoblot of ferritin in PMt, treated (lane +) or untreated (lane –) with Triton X -100, which were then subjected to trypsin d igestion. The intensity of the immuno- labeled band was lower in the presence of the detergent, demonstrating that ferritin is localized inside the mito- chondrial membranes. Furthermore, Mpl were obtained by osmotic shock of PMt to remove the outer mito- chondrial membrane. Mitoplast and PMt proteins were then cross-reacted with monoclonal antibodies raised Fig. 4. Ultrastructural localization of ferritin in pea stem mitochondria. Electron micrograph from fixed Percoll-purified pea st em mitochondria, subjected to immunogold decoration in th e presence (A and B) or absence (C) of PAAF and at lower magnification (D). Arrows indicate ele ctron- dense particles after immunolabeling with PAAF followe d by gold-conj ugated secondary ant ibody. Bars correspond to 300 lm. Ó FEBS 2004 Plant mitochondrial ferritin (Eur. J. Biochem. 271) 3661 against cytochrome c, polyclonal antibodies raised against the a/b-subunit of mitochondrial A TPase and PAAF, respectively (Fig. 2B). These results show that Mpl partially lost the cytochrome c; the densitometric a nalysis of the immunodecoration show a d ecrease of approximately 50% in the Mpl proteins. On the other hand, the immunodec- oration of PMt and Mpl proteins with antibodies against the a/b-subunit and PAAF was comparable (Fig. 2B). This indicates that both the a/b-subunit and the ferritin are still retained in Mpl, thus confirming the localization of ferritin in the mitochondrial matrix. The ultrastructural localization of the pea stem mito- chondrial ferritin was further confirmed in ultra-thin sections of pea s tem c ross-sections (Fig. 3) and fixed mitochondria (Fig. 4), both immunolabeled with PAAF, and followed by gold-conjugated secondary antibody. Figure 3 shows that mitochondria, selected from a trans- mission electron microscopic micrograph of pea stem cells (Fig. 3 A), have some electron-dense particles (Fig. 3B,C; arrows). In agreement, such particles were also detected in isolated mitochondria (Fig. 4A,B). The electron-dense par- ticles were not detected when PAAF was omitted in both cross-sections (result not shown ) and isolated mitochondria (Fig. 4C). According to the low level of plastidial enzymatic marker, detected in PMt (Table 1), the elec tron micro- graphs show that the purified mitochondrial fraction was almost free from etioplast contamination (Fig. 4D). Figure 5A shows the protein patterns of control proteins and o f CMt and PMt from A. thaliana.Whenthese separated proteins were subjected to cross-reaction with PAAF (Fig. 5B), a gain a band with an apparent molecular mass of 25– 26 kDa w as revealed, thus s uggesting the presence of this iron-storage protein also in mitochondria from this type of plant cells. Furthermore, preliminary results indicate that ferritin was also present in Percoll- purified mitochondria isolated from soybean hypocotyls (results not shown). The genome of A. thaliana contains a family of nuclear genes for ferritins (AtFer1–4) [28]. These genes encode the ferritin subunit precursors, each containing a transit peptide. The structural analysis of the presequences of the corres- ponding polypeptides suggests that all are targeted to plastids [28]. Table 3 shows the scores for t he mitochondrial/plasti- dial localization of some plant ferritins from P. sativum (SwissProt accession P19975), cowpea (Vigna unguiculata, SwissProt accession T08124), soybean (Glycine max, SwissProt accession BAB64536) and AtFer1 and AtFer4 from A. thaliana. While it is clear that AtFer1 is a poor candidate for a mitochondrial localization, for the other proteins significant scores were found. In particular, PSORT and IPSORT programs predicted high probability for the presence of a mitochondrial target peptide in pea ferritin. Remarkably, the ferritins from cowpea, soybean and AtFer4 exhibit values corresponding to a high probability for a mitochondrial targeting from at least three programs. Discussion Animal and plant ferritins are encoded by nuclear gene families, which diverge in their exon/intron organization [13]. This suggests that they derive from a common ancestor, albeit animal ferritins display a cytoplasmic localization, whereas the plant ones are plastidic [8,15]. However, as seen, an unusual intronless gene on human chromosome 5q23.1 encodes a 242 amino acid precursor of a ferritin H-like Fig. 5. Identification of ferritin in A. thaliana mitochondria. (A) SDS/ PAGE (12%) of proteins (25 lg) from crude mitochondria (CMt), purified mitochondria (PMt), and total pea seed proteins (CP, control proteins). (B) Immunoblotting of the same proteins with PAAF. Molecular mass of protein standards is indicated in kDa (Std). Table 3. Calculated values for prediction of mitochondrial targeting for some plant ferritins. Scores were obtaine d from different pro grams available on the net; values higher than 0.6 are highlighted in bold; the output for IPSORT is given as mitochondrial target peptide (mTP) or chloroplast transit peptide (cTP). Plant species Prediction programs SwissProt accession PREDOTAR MITOPROT II PSORT IPSORT Pisum sativum P19975 0.031 0.3601 0.611 mTP Vigna unguiculata T08124 0.783 0.9613 0.473 mTP Glycine max BAB64536 0.648 0.8187 0.617 mTP Arabidopsis thaliana Q39101 (AtFer1) 0.020 0.5482 0.360 cTP Q9S756 (AtFer4) 0.974 0.8137 0.694 mTP 3662 M. Zancani et al. (Eur. J. Biochem. 271) Ó FEBS 2004 protein [10]. This 30 kDa protein is targeted to mitochon- dria and processed to a 22 kDa subunit. This ferritin, expressed in HeLa cells, is available to exogenous iron, similarly to t he cytosolic ferritin, suggesting that t his mitochondrial protein may have profound consequences on c ellular i ron homeostasis [11]. A s shown in this p aper, a ferritin has also been identified in higher plant mitochon- dria. This evidence arises from the following findings. First, a mitochondrial matrix protein of 25–26 kDa, cross-reacted with a polyclonal antibody of pea seed ferritin in b oth pea stem (Fig. 1) and A. thaliana (Fig. 5) mitochondria. Such organelles were highly purified by discontinuous Percoll gradient, providing a v ery low interference in the immuno- decoration from other cellular c omponents, especially from etioplast contamination. On the basis of densitometric analysis of immunoblots obtained with etioplasts isolated from pea stem (results not shown), we calculated that if PAAF detects j ust etioplast ferritin, these organelles h ave to be present, in purified mitochondrial fractions, as a heavy contamination (estimated to be approximately 25% of the total protein). The results shown here demonstrate that this is not the case, because the low enzymatic activity of glucose-6-phosphate dehydrogenase in PMt (Table 1) con- firms that the purified mitochondrial fractions possess a maximum of 2 .5% of plastid proteins and, in addition , the electron micrographs (Fig. 4) clearly show a very limited contamination of P Mt from other o rganelles. Second, ferritin was immunocytochemically identified in etiolated pea stem cross-sections (Fig. 3) and in isolated pea mito- chondria (Fig. 4). The pea stem mitochondrial ferritin is present in the mitochondrial matrix as demonstrated by its colocalization in Mpl with the a/b-subunit of mitochondrial ATPase (Fig. 2). Finally, the 25–26 kDa soluble protein was purified by immunoprecipitation (Fig. 1D); the primary structure o f the polypeptide chain, inferred by Mass Finger Print experiments on MALDI-TOF mass spectrometry, fits to a high degree with the sequence of the ferritin from P. sativum (SwissProt accession P19975, Table 2). In A. thaliana, four ferritin genes (AtFer1–4) have been reported and it has been suggested that the proteins AtFer1– 4 possess at the N-terminus the typical presequences of the chloroplastic protein transit peptide [28], similarly to what reported for pea ferritin [29,30]. On the other hand, the analysis of the presequence of AtFer4 reveals a high score for its mitochondrial localization, especially when co mpared with AtFer1 (Table 3). The same analysis for pea ferritin shows that the programs PSORT and IPSORT give a high probability for this protein to be targeted to mitochondria (Table 3). The data presented in this paper strongly indicate a mitochondrial localization for ferritins in P. sativum and A. thaliana and could be rationalized as follows: the protein may be targeted to both plastids and mitochondria, similarly to what shown for several plant p roteins [31]. This feature can be accomplished by alternative transcription, alternative translation starts, alternative exon splicing (or a combination of the above), or t he presence in the N-terminus of an ambiguous presequence [31]; prediction programs could just be unable to detect such dual targeting. On the other hand, a similar situation has been described for ferrochelatase-I, an enzyme involved in heme biosyn- thesis and, probably, in protection against o xidative stress in A. thaliana [32]. This enzyme has been recently rep orted to be present also in pea mitochondria [33]. The presence of ferrochelatase-I and -III in A. thaliana mitochondria has been recently questioned, while their presence in pea mitochondria has been related to the fact that the latter organelles import a variety of (but not all) chloroplastic proteins [34]. Plant mitochondria possess an electron transport chain where superoxide anion may be generated by univalent reactions at the level of complex I or III [35]. For this reason, mitochondria have evolved systems to scavenge ROS, or to prevent their formation [7,9], but sequestration of potential harmful ferrous ions has not yet been described. Metal tolerance and homeostasis in plant cells is accom- plished by different mechanisms [36]. In this context, the main role of ferritins could concern iron sequestration. Overexpression of this protein, in either the cytoplasm or plastids of transgenic tobacco, leads to an increase of iron sequestration that induces an activation of the iron trans- port systems [37]. Therefore, they are crucial in controlling iron storage and homeostasis in the plant cells. Other functions of plant ferritins are, on the other hand, still obscure. It has been suggested that sequestering of intracel- lular iron m ay protect from oxidative damage induced by a wide range of stresses [38]. Indeed, an increase of ferritin mRNA has been observed in A. thaliana leaves photo- inhibited w ith high light or fumigated with ozone [39]. Therefore, the sequestration of iron by ferritins in chloro- plasts and mitochondria, two of the major sites of ROS generation in plant cells [6,7], can constitute an additional strategy to prevent th is damage. Acknowledgements We thank Dr J.F. Briat, Centre National d e l a R echerche S cientifique, Montpellier, France, for a gene rous gift of pea s eed ferritin antibody. We also thank very much Dr Sonia Levi, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS), H. San Raffaele, Milan, Italy, for the cross-reactivity analysis with human mitochondrial ferritin antibo dy. Thanks are also due to Dr Sonia Patui for her help during Percoll- purified mito chondria preparation and to Mr Claudio Gamboz for his help with electron microscopy analysis. This research was s upported by ‘Ministero dell’Universita ` e della Ricerca Scientifica e Tecnologica’ (Cofin 2000–01) i n the fr ame of the program entitled: Nitric Oxide an d Plant Resistance to Pathogens. References 1. Briat, J F., Fo bis-Loisy, I., Grignon, N., Lobre ´ aux, S., Pascal, N., Savino, G., Thoiron, S., von Wire ` n, N . & Van Wuytswinkel, O. (1995) Cellular and molecular aspects of iron metabolism in plants. Biol. Cell. 84, 69–81. 2. Halliwell, B. & Gutteridge, J.M.C. (1999) Free Radicals in Biology and Medicine, 3rd edn. Oxford University Press, Oxford. 3. Elstner, E.F. & Osswald, W. (1994) Mechanisms of oxygen activation during plant stress. Proc. Royal Soc. Edinburgh 102B, 131–154. 4. Mehdy, M.C., Sharma, Y.K., Sathasivan, K. & Bays, N.W. (1996) The role of activated oxygen species in plant disease resistance. Physiol. Plant. 98, 3 65–374. 5. Briat, J F. & Lobre ´ aux, S. (1997) Iron transport and s torage in plants. Trends Plant Sci. 2, 187–192. Ó FEBS 2004 Plant mitochondrial ferritin (Eur. J. Biochem. 271) 3663 6. Alscher, R.G., Donahue, J.N. & Cramer, C.L. (1997) Reactive oxygen species and antioxidants: relationship in green cells. Phy- siol. Plant. 100, 224–233. 7. Møller, I.M. (2001) Plant mitochondria and oxidative stress: electron transport, NADPH turnover, and me tabolism of reactive oxygen species. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52 , 561– 591. 8. Seckbach, S. (1982) Ferreting out the secrets of plant ferritin. J. Plant. Nutr. 5, 369–394. 9. Casolo, V., Braidot, E., Chiandussi, E ., Macrı ` ,F.&Vianello,A. (2000) The role of mild uncoupling and non-coupled respiration in the regulation of hydrogen peroxide generation by plant mitochondria. FEBS Lett. 474, 53–57. 10. Levi, S., Corsi, B., Bosisio, M., Invernizzi, R., Volz, A., Sanford, D., Arosio, P. & Drysdale, J. (2001) A human mitochondrial ferritin encoded by an introneless gene. J. Biol. Chem. 270, 24437–24440 11. Corsi, B., Cozzi, A., Arosio, P., Drysdale, J., Santambrogio, P., Campanella, A., Biasiotto, G., Alberini, A. & Levi, S. (2002) Human mitochondrial fe rrit in expressed in HeLa cells incorporates iron and affects cellular iron metabolism. J. Biol. Chem. 277, 22430–22437. 12. Theil, E.C. (1987) Ferritin: structure, gene regulation, and cellular function in animals, plants, a nd microorganisms. Ann. Rev. Biochem. 56, 289–315. 13. Harrison, P.M. & Arosio, P. (1996) The ferritins: molecular properties, iron storage function and cellular regulation. Biochim. Biophys. Acta 1275, 161–203. 14. Laulhere, J.P., Laboure, A M. & Briat, J F. (1989) Mechansim of the transition from plant ferritin to phytosiderin. J. Biol. Chem. 264, 3629–3635. 15. Proudhon, D., Wei, J., Briat, J.F. & Theil, E.C. (1996) Ferritin gene organization: differenc es between plants and animals s uggest possible kingdom-specific selective constraints. J. Mol. Evol. 42, 325–336. 16. Zancani, M., Casolo, V., Vianello, A. & Macrı ` ,F.(1998)H + /PP i stoichiometry of a membrane-bound pyrophosphatase of plant mitochondria. Physiol. Plant. 103, 304–311. 17. Petrussa,E.,Casolo,V.,Braidot,E.,Chiandussi,E.,Macrı ` ,F.& Vianello, A. (2001) Cyclosporin A induces the opening of a potassium-selective channel in higher plant mitochondria. J. Bioenerg. Biomembr. 33, 107–117. 18. Macrı ` , F., Braidot, E., Petrussa, E. & Vianello, A. (1994) L ipoxy- genase activity associated to isolated soybean plasma membranes. Biochim. Biophys. Acta 1215, 109–114. 19. Green, J.R. (1983) The Golgi apparatus. In Isolation of Mem- branes and Organelles from Plant Cells (Hall, J.L. & Moore, A.L., eds), pp. 135–152. Acad emic Press Inc., London. 20. Lord, J.M. (1983) Endoplasmic reticulum and ribosomes. In Isolation of Membranes and O rganelles from Plant Cells (Hall, J.L. & Moore, A.L., eds), pp. 119–134. Academic Press Inc., London. 21. Bergmeyer, H.U., Gawehn, K. & Grassl, M. (1974) Enzymes as biochemical reagents; glucose-6-phosphate dehydrogenase. In Methods of Enzymatic Analysis,Vol.1(Bergmeyer,H.U.,ed.),pp. 458–459. Academic Press Inc., London. 22. Laulhere, J.P., Lescure, A.M. & Briat, J.F. (1988) Purification a nd characterization of ferritins from seed of maize, pea and soya bean: distribution in various pea organs. J. Biol. Chem. 262, 10289–10294. 23. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 277, 680–685. 24. Tomasetig, L., Di P ancrazio, F., Harris, D.A., Mavelli, I. & Lippe, G. (2002) Dimerization of F o F 1 ATP synthase from bovine heart is independent from the binding of the inhibitor p rotein IF 1 . Biochim. Biophys. Acta 1556, 133–141. 25. Shevchenko, A. & Shevchenko, A. (2001) Evaluation of the effi- ciency of in-gel digestion of p roteins by peptide isotopic labelling and MALDI ma ss spe ctro metry . Anal. Bio chem. 296 , 279–283. 26. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye bindin g. Anal. Biochem. 72, 248–254. 27. Møller, I.M. & Lin, W. (1986) Membrane-bound NAD(P)H dehydrogenases in higher plant c ells. Ann. Rev. Plant Physiol. 37, 309–334. 28. Petit, J.M., Briat, J F. & Lobreaux, S. (2001) Structure and dif- ferential expression of the four members of the Arabidopsis thali- ana ferritin gene family. Biochem. J. 359, 575–582. 29. Lobre ´ aux, S., Yewdall, S.J., Briat, J F. & Harrison, P.M. (1992) Amino-acid sequence and predicted three-dim ensional structure of pea seed (Pisum sativum)ferritin.Biochem. J. 288, 931–939. 30. Van Wuytswinkel, O., Savino & Briat, J F. (1995) Purification and characterization o f re combinant pe a-seed ferritins expressed in Escherichia coli: in fluence of N-terminus deletions on protein solubility and c ore formation in vitro. Biochem. J . 305 , 253–261. 31. Peeters, N. & Small, I. (2001) Dual targeting to mitochondria and chloroplasts. Biochim. Biophys. Acta 1541, 54–63. 32. Chow,K.S.,Singh,D.P.,Roper,J.M.&Smith,A.G.(1997)A single precursor protein for ferrochelatase-I from Arabidopsis is imported in vitro into both chloroplasts and mitochondria. J. Biol. Chem. 272, 27565–27571. 33. Cornah, J.E., Roper, J.M., Pal Singh, D. & Smith, A.G. (2002) Measurement of ferrochelatase activity using a novel assay sug- gests that plastids a re the major site of haem biosynthesis in both photosynthetic and non-photosynthetic cells of pea (Pisum sati- vum L.). Biochem. J. 362, 423–432. 34. Lister, R., Chew, O., Rudhe, C., Lee, M N. & Whelan, J. (2001) Arabidopsis thaliana ferrochelatase-I an d -II are not impo rted into Arabidopsis mitochon dria. FEBS Lett. 506, 291–295. 35. Braidot, E., Petrussa, E., Vianello, A. & Macrı ` , F. (1999) Hydrogen peroxide generation by higher plant mitochondria oxidizing complex I or complex II substrates. FEBS Lett. 451, 347–350. 36. Clemens, S. (2001) Molecular mechanisms of plant metal tolerance and homeostasis. Planta 212, 475–486. 37. Van Wuytswinkel, O., Vansuyt, G., Grignon , N., Fourcroy, P. & Briat, J F. (1998) Iron homeostasis alteration in transgenic tobacco overexpressin g ferritin. Plant J. 17, 93–97. 38. Dea ´ k, M., Horva ´ th, G.V., Davletova, S., To ¨ ro ¨ k, K., Sass, L., Vass, I., Barna, B., Kira ´ ly,Z.&Dudits,D.(1999)Plantsectopi- cally expressing the iron-binding p rotein, ferritin, are tolerant t o oxidative damage and pathogens. Nature Biotechnol. 17, 192–196. 39. Murgia, I., Briat, J F., Tarantino, D. & Soave, C. (2001) Plant ferritin accumulates in response to photoinhibition but its ectopic overespression does not protect against photoinhibition. Plant Physiol. Biochem. 39, 797–705. 40. Vianello, A., Dell’Antone, P. & Macrı ` , F. (1982) ATP-dependent and ionophore-ind uced pr oton translo cation in pe a stem micr o- somal vesicles. Biochim. Biophys. Acta 689, 89–96. 41. Eichacker, L.A., Helfrich, M., Ru ¨ diger, W. & Mu ¨ ller, B. (1996) Stabilization of chlorophyll a-binding apoprotein s P700, CP47, CP43, D2, and D1 by chlorophyll a or Zn-pheophytin a. J. Biol. Chem. 271, 32174–32179. 3664 M. Zancani et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . 2B). This indicates that both the a/b-subunit and the ferritin are still retained in Mpl, thus confirming the localization of ferritin in the mitochondrial. element for all living organisms [1]. In green plants, its importance mainly derives from the presence at the active sites of metalloproteins involved in the