Identification and Functional Characterization of a Novel Mitochondrial Carrier for Citrate and Oxoglutarate in Saccharomyces cerevisiae*

Mitochondrial carriers are a family of transport proteins that shuttle metabolites, nucleotides, and coenzymes across the mitochondrial membrane. The function of only a few of the 35 Saccharomyces cerevisiae mitochondrial carriers still remains to be uncovered. In this study, we have functionally defined and characterized the S. cerevisiae mitochondrial carrier Yhm2p. The YHM2 gene was overexpressed in S. cerevisiae, and its product was purified and reconstituted into liposomes. Its transport properties, kinetic parameters, and targeting to mitochondria show that Yhm2p is a mitochondrial transporter for citrate and oxoglutarate. Reconstituted Yhm2p also transported oxaloacetate, succinate, and fumarate to a lesser extent, but virtually not malate and isocitrate. Yhm2p catalyzed only a counter-exchange transport that was saturable and inhibited by sulfhydryl-blocking reagents but not by 1,2,3-benzenetricarboxylate (a powerful inhibitor of the citrate/malate carrier). The physiological role of Yhm2p is to increase the NADPH reducing power in the cytosol (required for biosynthetic and antioxidant reactions) and probably to act as a key component of the citrate-oxoglutarate NADPH redox shuttle between mitochondria and cytosol. This protein function is based on observations documenting a decrease in the NADPH/NADP+ and GSH/GSSG ratios in the cytosol of ΔYHM2 cells as well as an increase in the NADPH/NADP+ ratio in their mitochondria compared with wild-type cells. Our proposal is also supported by the growth defect displayed by the ΔYHM2 strain and more so by the ΔYHM2ΔZWF1 strain upon H2O2 exposure, implying that Yhm2p has an antioxidant function.

Yhm2p is a protein of unknown function that has been found to be present in the inner mitochondrial membrane of Saccharomyces cerevisiae (1). This protein was shown to complement the growth defect displayed at 37°C by S. cerevisiae cells lacking ABF2, which encodes a mitochondrial DNA-binding protein involved in mitochondrial DNA replication and recombination (1,2). In addition, the primary structure of Yhm2p exhibits all the characteristic features of the mitochondrial carrier (MC) 3 protein family, i.e. a tripartite structure consisting of three tandemly repeated homologous domains of about 100 amino acids in length, each of which contains a distinct signature motif and two hydrophobic stretches that span the membrane as ␣-helices (for reviews see Refs. [3][4][5]. S. cerevisiae possesses 35 members of this family, one of which is localized in peroxisomes and not in mitochondria. The majority of yeast MCs has been functionally characterized and shown to transport specific metabolites, nucleotides, and coenzymes across the mitochondrial membrane (for review see Ref. 6). By contrast, the substrate(s) transported by Yhm2p have not yet been discovered, and its physiological function is yet unknown.
In this study, we provide evidence that the gene product of YMR241w, named Coc1p and known as Yhm2p, is a citrateoxoglutarate carrier in S. cerevisiae. Yhm2p was overexpressed in S. cerevisiae, purified, reconstituted into phospholipid vesicles, and identified for its transport properties as a novel carrier for citrate and oxoglutarate. Besides these substrates, oxaloacetate, succinate, and fumarate are also transported, to a lesser extent, by the carrier via a counter-exchange mechanism. Yhm2p operates in yeast mitochondria with transport properties similar to those observed with the recombinant protein. In addition, ⌬YHM2 cells (devoid of YHM2) and more so ⌬YHM2⌬ZWF1 cells (lacking YHM2 and also ZWF1 which encodes glucose-6-phosphate dehydrogenase, an important source of NADPH) exhibit a growth defect upon exposure to H 2 O 2 . This phenotype correlates with a decrease in the cytosolic NADPH/NADP ϩ and GSH/GSSG ratios. This is the first time that an MC for both citrate and oxoglutarate has been identified from any organism at the molecular level. Yhm2p is suggested to have a role in increasing NADPH levels in the cytosol and to be a key component of the * This work was supported by grants from the Ministero dell'Università e della NADPH citrate-oxoglutarate redox shuttle that catalyzes a net transfer of NADPH reducing equivalents from the mitochondria to the cytosol.

Yeast Strains, Media, and Preparation of Mitochondria and
Cytosol-The single and double deletion strains of YPH499 (wild type) were constructed using the PCR-mediated gene disruption technique by replacing the open reading frame of the genes under investigation with HIS3 (7), KanMX3 (8), or NatMX4 (9) (see supplemental Table S1). All deletions were verified by PCR. The wild-type and deletion strains were grown at 30°C in rich medium (YP), containing 2% bactopeptone and 1% yeast extract, synthetic complete medium (SC), or synthetic minimal medium (SM) (10) supplemented with 2% glucose or 2% acetate and auxotrophic nutrients when required. The final pH was adjusted to 4.5 (with glucose) or 6.5 (with acetate). When using solid media, 1.5% agar was added. In nearly all the experiments, yeast strains were precultured overnight in YP with 2% glucose. After washing, the cells were suspended in acetate-supplemented SC at an absorbance of 0.1 at 600 nm and grown until the early exponential phase was reached (absorbance of 0.8). These cells were incubated in the presence or absence of H 2 O 2 (1 mol ϫ 10 7 cells) for 60 min. At this time period, the cells were harvested and centrifuged at 2500 ϫ g for 10 min at 4°C. For the preparation of the mitochondria, the pellet was resuspended in 1.2 M sorbitol and 20 mM potassium phosphate, pH 7.4, and the mitochondria were isolated as described previously (11). To prepare the cytosol, the pellet was resuspended in 0.65 M sorbitol, 0.1 M Tris-HCl, pH 8.0, and 5 mM EDTA. The cells were then disrupted by employing a cell disrupter (model Z6/40/AE/GA, Constant Systems Ltd., Daventry, UK) set at a working pressure of 207 MPa. Cellular lysates were centrifuged at 2500 ϫ g for 10 min at 4°C and then at 120,000 ϫ g for 30 min at 4°C to obtain a clear supernatant (cytosol).
Reconstitution into Liposomes and Transport Assays-Purified YHM2 protein was reconstituted into liposomes in the presence or absence of substrates, as described previously (12). External substrate was removed from proteoliposomes on Sephadex G-75 columns pre-equilibrated with 50 mM NaCl and 10 mM PIPES, pH 7.0 (buffer G). Transport at 25°C was started by adding [ 14 C]citrate or [ 14 C]oxoglutarate to proteoliposomes and terminated by the addition of 30 mM pyridoxal 5Ј-phosphate and 10 mM bathophenanthroline, which in combination and at high concentrations inhibit the activity of several mitochondrial carriers completely and rapidly (see, for example, Refs. [13][14][15][16]. In controls, the inhibitors were added with the labeled substrate. The external radioactive substrate was removed, and the radioactivity in the proteoliposomes was measured (12). The experimental values were corrected by subtracting control values. The initial transport rate was calculated from the radioactivity taken up by proteoliposomes after 1 min (in the initial linear range of substrate uptake). For efflux measurements, proteoliposomes containing 1 mM substrate were labeled with 10 M [ 14 C]citrate by carrier-mediated exchange equilibration (12). After 60 min, the external radioactivity was removed by passing the proteoliposomes through Sephadex G-75 pre-equilibrated with buffer G. Efflux was started by adding unlabeled external substrate in buffer G, or buffer G alone, and terminated by adding the inhibitors indicated above.
Metabolite, Pyrimidine Coenzyme, and Glutathione Determinations by Mass Spectrometry-Mitochondria and cytosol were extracted with phenol/chloroform 1:1 for mass spectrometry analysis of NADP ϩ , NADPH, NAD ϩ , NADH, reduced and oxidized glutathione, and with NH 3 -saturated phenol/chloroform 1:1 for analysis of citrate and oxoglutarate. A Quattro Premier mass spectrometer interfaced with an Acquity UPLC system (Waters) was used for liquid chromatography-tandem mass spectrometry analysis. Calibration curves were established using standards, processed under the same conditions as the samples, at five concentrations. The best fit was determined using regression analysis of the peak analyte area.  16 for GSSG. Chromatographic resolution was achieved using HSS T3 (2.1 ϫ 100 mm, 1.8-m particle size, Waters) for citrate and oxoglutarate; Atlantis dC18 (4.6 ϫ 150 mm, 5 m particle size, Waters) for pyrimidine coenzymes; and BEH C18 (2.1 ϫ 50 mm, 1.7 m particle size, Interchim) for reduced and oxidized glutathione. For all columns, the flow rate was 0.3 ml/min. ROS Measurements-ROS were assessed using 2Ј,7Ј-dichlorodihydrofluorescein diacetate (DCFH-DA) (17,18). Fifteen micromolar DCFH-DA was added to cells in acetate-supplemented SC at the early exponential phase; after 30 min, the cells were washed once. For flow cytometric analysis, DCFH-DApreloaded cells were incubated in the presence and absence of H 2 O 2 (1 mol ϫ 10 7 cells) for the indicated times at 30°C; DCF fluorescence was measured using a Coulter Epics Elite ESP (Beckman-Coulter) equipped with a 15-milliwatt argon-ion laser (excitation wavelength, 488 nm; emission wavelength, 520 nm) (19). For microplate reader assay, DCFH-DA-preloaded cells were incubated in the presence and absence of H 2 O 2 (0.1 mol ϫ 10 5 cells) for the indicated times at 30°C; DCF fluorescence was measured at 485-nm excitation and 538-nm emission wavelength (20).
Other Methods-Proteins were separated by SDS-PAGE and either stained with Coomassie Blue dye or transferred to nitrocellulose membranes. Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry of trypsin digests of the purified Yhm2p band excised from a Coomassiestained gel was carried out as described (21). Western blotting was carried out with rabbit antiserum against the bacterially expressed Yhm2p protein. The amount of purified protein was estimated by laser densitometry of stained samples using carbonic anhydrase as a protein standard (22). The amount of protein incorporated into liposomes was measured as described (22) and was ϳ20% of the protein added to the reconstitution mixture. To quantify the percentage of nonviable cells, wildtype and deleted strains in acetate-supplemented SC at the early exponential phase were exposed to H 2 O 2 (1 mol ϫ 10 7 cells) for 1 h at 30°C. Then they were washed in 15 mM MgCl 2 and 50 mM Tris-HCl, pH 7.7, resuspended in the same buffer containing 0.046 mM propidium iodide, and analyzed with a Coulter Epics Elite ESP (Beckman-Coulter) (23).

Expression in S. cerevisiae and Purification of Yhm2p-
Yhm2p was expressed in S. cerevisiae from the multicopy pYES2 vector and purified by nickel-nitrilotriacetic acid-agarose chromatography (supplemental Fig. S1). Its apparent molecular mass was about 34 kDa (supplemental Fig. S1A, Functional Characterization of Recombinant Yhm2p-In the search for potential substrates of Yhm2p, we based our choice on the fact that this protein displays residues at the proposed common substrate-binding site of MCs (24) that are nearly identical to those of the rat CTP and its yeast ortholog Ctp1p, which are known to transport citrate in exchange for malate (25)(26)(27)(28). Thus, the substrate binding points are as follows: Gly 93 -Lys 97 -Arg 101 (point I), Lys 190 -Gln 191 (point II), and Arg 285 (point III) for CTP; Gly 79 -Lys 83 -Arg 87 (point I), Arg 181 -Gln 182 (point II), and Arg 279 (point III) for Ctp1p, which correspond to Glu 83 -Lys 87 -Leu 91 (point I), Arg 181 -Gln 182 (point II), and Arg 279 (point III) in Yhm2p. Therefore, recombinant and purified Yhm2p was reconstituted into liposomes, and its ability to transport citrate was tested. Indeed, proteoliposomes reconstituted with Yhm2p catalyzed active [ 14 C]citrate/citrate exchange, which was abolished by a mixture of the inhibitors pyridoxal 5Ј-phosphate and bathophenanthroline. They did not catalyze homo-exchanges for phosphate, aspartate, glutamate, ornithine, ADP, ATP, and GTP (internal concentration, 10 mM; external concentration, 1 mM). Furthermore, no [ 14 C]citrate/citrate exchange activity was detected if recombinant Yhm2p had been boiled before incorporation into liposomes.
The substrate specificity of purified and reconstituted Yhm2p was investigated in detail by measuring the uptake of [ 14 C]citrate into proteoliposomes that had been preloaded with various potential substrates (Fig. 1A). The highest rates of [ 14 C]citrate uptake into proteoliposomes were found with internal citrate, oxoglutarate, oxaloacetate, succinate, and fumarate. Surprisingly, at variance with the known mitochondrial citrate/malate carrier (25)(26)(27)(28), reconstituted Yhm2p transported isocitrate, cis-aconitate, and L-malate with very low efficiency; it transported trans-aconitate at the same rate as cis-aconitate but did not transport 1,2,3-pentanetricarboxylate, maleate, and malonate. Yhm2p did also not catalyze uptake of [ 14 C]citrate in exchange for internal glutamate, oxoadipate, aspartate, phosphate, sulfate, and pyruvate and not shown ADP, ATP, GTP, thiosulfate, ornithine, glutamine, and carnitine.
The sensitivity of the Yhm2p-mediated [ 14 C]citrate/citrate exchange reaction by known inhibitors of different MCs was also examined. Thiol reagents (mersalyl, p-chloromercuribenzenesulfonate, mercuric chloride, and N-ethylmaleimide), the lysine reagent pyridoxal 5Ј-phosphate as well as bathophenanthroline and tannic acid (whose mechanism of action is not known) strongly inhibited citrate exchange; ␣-cyanocinnamate and bromcresol purple (whose mechanism of action is also unknown) inhibited citrate exchange partially (Fig. 1B). In contrast, 1,2,3-benzenetricarboxylate and carboxyatractyloside, the powerful inhibitors of the citrate/malate carrier (27) and the ADP/ATP carrier (29), respectively, had very little effect on the activity of Yhm2p. The competitive inhibitors of the dicarboxylate and oxoglutarate/malate carriers, butylmalonate and phenylsuccinate, were also rather ineffective. In addition, externally added citrate, oxoglutarate, succinate, and fumarate inhibited the [ 14 C]citrate/citrate exchange (Fig. 1B). A low inhibition was found with malate and virtually no effect was detected with glutamate and ATP.
Kinetic Characteristics of Recombinant Yhm2p-The time courses of [ 14 C]citrate and [ 14 C]oxoglutarate uptake into proteoliposomes were measured either as uniport (in the absence of internal substrate) or as antiport (in the presence of internal citrate or oxoglutarate, respectively) ( Fig. 2A). The uptake of citrate or oxoglutarate by exchange followed a first order kinetics (rate constants, 0.18 and 0.13 min Ϫ1 ; initial rates, 8.84 and 6.29 mmol/min/g of protein for citrate and oxoglutarate, respectively), isotopic equilibrium being approached exponentially. In contrast, the uniport uptake of citrate and oxoglutarate was negligible, suggesting that Yhm2p does not catalyze a unidirectional transport (uniport) of citrate or oxoglutarate. This issue was further investigated by measuring the efflux of  [ 14 C]citrate from prelabeled active proteoliposomes, as this provides a more sensitive assay for unidirectional transport (12). In the absence of external substrate, no efflux was observed even after a 30-min incubation (Fig. 2B). However, addition of external citrate or oxoglutarate led to extensive efflux of radioactivity, which was prevented completely by the presence of the inhibitors pyridoxal 5Ј-phosphate and bathophenanthroline (Fig. 2B). These results demonstrate that, at least under the experimental conditions employed, reconstituted Yhm2p catalyzes an obligatory antiport reaction of substrates.
The kinetic constants of purified reconstituted Yhm2p were determined by measuring the initial transport rate at various external [ 14  Impaired Citrate Uptake in Mitochondria Lacking Yhm2p-In other experiments, the uptake of [ 14 C]citrate was measured in proteoliposomes that had been reconstituted with Triton X-100 extracts of mitochondria isolated from wild-type, ⌬YHM2, and YHM2-pYES2 cells (Fig. 3). A low citrate/citrate exchange activity was observed upon reconstitution of mitochondrial extracts from the knock-out strain, probably because of the presence of Ctp1p. This activity more than doubled in parental mitochondrial extracts and was even greater in mitochondrial extracts from YHM2-pYES2 cells. Similar results were obtained by measuring the uptake of [ 14 C]citrate in exchange with intraliposomal oxoglutarate, succinate, fumarate, and oxaloacetate (Fig. 3), which are substrates of Yhm2p, but not with substrates of other MCs such as phosphate, ADP, maleate, 1,2,3-pentanetricarboxylate, glutamate, carnitine, and ornithine. Moreover, the citrate/citrate exchange in proteoliposomes reconstituted with mitochondrial extracts from wildtype and YHM2-pYES2 cells was markedly inhibited by mercurials, N-ethylmaleimide and tannic acid (data not shown). As a control, the [ 14 C]ADP/ATP exchange (the defining reaction of the ADP/ATP translocase) and the [ 14 C]malate/phosphate exchange (the defining reaction of the dicarboxylate carrier) were virtually the same for all types of reconstituted mitochondrial extracts (data not shown). Moreover, as demonstrated by immunodecorating wild-type and ⌬YHM2 mitochondria with specific antibodies, the amounts of the ADP/ATP, phosphate, and dicarboxylate carriers were virtually the same (results not shown). Therefore, in mitochondria Yhm2p exhibits the same specificity and inhibitor sensitivity as the purified protein; moreover, the absence of YHM2 does not affect the expression or the activity of other MCs.
Subcellular Localization of Yhm2p-Yhm2p has been reported to be localized to the inner mitochondrial membrane (1). However, these studies did not exclude the possibility that Yhm2p is also localized to peroxisomes, notably because its transport properties make it a candidate for being involved in a peroxisomal redox shuttle that is required for the intraperoxisomal regeneration of NADPH (30,31). Also, Yhm2p was among the proteins that had been identified in a purified peroxisomal membrane fraction by mass spectrometry (32). The subcellular distribution of Yhm2p was therefore analyzed by separating a post-nuclear supernatant of oleic acid-induced wild-type cells on a sucrose density gradient. The enzyme activities of the mitochondrial marker enzyme fumarase and peroxisomal enzyme catalase were separated with some overlapping in fractions 13-17 (supplemental Fig. S2A). The immunoblots of fractions 3-29 demonstrate co-localization of Yhm2p with the mitochondrial phosphate transporter Mir1p and no co-localization of Yhm2p with Fox3p, although overlapping was observed in fraction 15 (supplemental Fig. S2A). To obtain further proof of the localization of Yhm2p, BJ1991pox1⌬, the strain used for the mass spectrometric analysis of peroxisomal membrane proteins (32), was induced in the presence of oleic acid. A post-nuclear supernatant was separated on a sucrose gradient, and the fractions enriched for peroxisomes were pooled, concentrated, and further separated on an Accudenz gradient. The subsequent immunological analysis using anti-Fox3p and anti-Pex11p antibodies showed peroxisomes peaking in fractions 8 and 9, whereas mitochondrial Mir1p peaked in fractions 12-15 (supplemental Fig. S2B) demonstrating the successful separation of peroxisomes from mitochondria. The Yhm2p-specific signal co-migrated with that of Mir1p, and no additional peroxisomal peak was detectable (supplemental Fig. S2B).
Localization of Yhm2p was also investigated in live cells by expressing a Yhm2p-GFP fusion protein from the native YHM2 promoter. Co-expression of a synthetic mitochondrial inner membrane marker protein (PrF 0 ATP9) that had been fused with DsRed (33) resulted in an overlapping pattern of fluorescence for the two proteins (data not shown). Contrarily, coexpression of a peroxisomally targeted PTS2-DsRed construct (33), displaying the typical punctate peroxisomal staining pattern, revealed that Yhm2p-GFP-dependent fluorescence did not coincide with peroxisomes (data not shown). These data demonstrate that Yhm2p is not a peroxisomal protein, in contrast to a previous conclusion (32), but is a mitochondrial protein in agreement with Cho et al. (1).
Growth Phenotypes of Knock-out Yeast Strains and Mitochondrial Citrate Levels-Having established the transport properties of Yhm2p by in vitro assays and its mitochondrial localization, the physiological role of Yhm2p was investigated. We first tested the ability of the ⌬YHM2 strain to grow under various conditions. Yeast cells lacking YHM2 exhibited substantial growth on YP and SC media containing either fermentative or nonfermentative carbon sources, similarly to the wildtype strain. However, on acetate-supplemented SM, ⌬YHM2 cells exhibited a slightly greater growth defect than that of ⌬ZWF1 cells (Fig. 4). The double mutant ⌬YHM2⌬ZWF1 exhibited a more pronounced defect than the single mutants when compared with wild-type cells. The growth of ⌬CTP1, ⌬CTP1⌬ZWF1, and ⌬YHM2⌬CTP1 cells was less affected (Fig.  4). On acetate-supplemented SC, in the presence of amino acids, none of the investigated strains showed any defect (Fig.  5A). However, when treated with H 2 O 2 , ⌬YHM2 cells exhibited a growth defect that was more evident in the double mutant ⌬YHM2⌬ZWF1 (Fig. 5A). Also, in the presence of H 2 O 2 , the ⌬CTP1 strain exhibited a lesser growth defect than the ⌬YHM2 strain as evidenced by the second and third dilutions of the two strains. Moreover, the ⌬CTP1⌬ZWF1 and ⌬YHM2⌬CTP1 double mutants were less sensitive to the H 2 O 2 -induced stress than the ⌬YHM2⌬ZWF1 double mutant; their growth defects were almost equal to those caused by the ⌬ZWF1 and ⌬YHM2 strains, respectively (Fig. 5A). We then investigated the effect of deleting YHM2 on citrate content in the mitochondria of S. cerevisiae cells grown on acetate-supplemented SC. ⌬YHM2 cells showed a significant increase (by 77%) in the amount of intramitochondrial citrate (1.6 Ϯ 0.2 nmol/mg of protein) as compared with wild-type cells (0.9 Ϯ 0.2 nmol/mg of protein), suggesting that Yhm2p may function to catalyze export of citrate from mitochondria to cytosol. In the cytosol, citrate is converted to isocitrate, which is the substrate of isocitrate dehydrogenase (Idp2p) that produces oxoglutarate and reduces NADP ϩ to NADPH. On this basis, the observed phenotypes could be due to diminished NADPH in the cytosol because this coenzyme is used to produce important antioxidants, such as GSH by glutathione reductase.
The first indication in support of this hypothesis stems from experiments carried out at 37°C in the presence or absence of phytosphingosine. When cultured at Ն37°C, yeast cells undergo major changes to increase thermotolerance and the ability to continue to divide. For example, at 37°C they increase the synthesis of ceramides by means of a pathway that includes  phytosphingosine as an intermediate (34,35). Given that NADPH is required in the last two reactions of phytosphingosine biosynthesis catalyzed by Tsc10 and Sur2, respectively, we tested the ability of our strains to grow at 37°C in acetatesupplemented SC. ⌬YHM2⌬ZWF1 cells were temperaturesensitive, and this phenotype was partially rescued by the addition of phytosphingosine (Fig. 5B). The deletant ⌬YHM2 also exhibited a slight growth defect at 37°C as evidenced by comparing the third and fourth dilutions between wild-type and ⌬YHM2 cells; this defect was also prevented by phytosphingosine (Fig. 5B). The double knock-out ⌬YHM2⌬CTP1 exhibited a temperature sensitivity almost equal to that of the ⌬YHM2 strain but not greater.
Cytosolic Levels of Nicotinamide Coenzymes and Glutathione-In light of the phenotypes exhibited by the knock-out strains for YHM2 and ZWF1 upon exposure to H 2 O 2 or at 37°C, the levels of cytosolic NADPH and NADP ϩ were measured. Without H 2 O 2 treatment, the cytosolic NADPH/NADP ϩ ratio in the ⌬YHM2 strain was significantly lower than that measured in the wild-type strain (Fig. 6A). This ratio decreased further (by about 43 and 48%) in the ⌬ZWF1 and ⌬YHM2⌬ZWF1 strains; however, it was not decreased in ⌬CTP1 cells as compared with wild-type cells. Moreover, no significant variations in the total amount of NADP ϩ plus NADPH were observed among the different strains (Fig. 6A). After treatment with H 2 O 2 , the NADPH/NADP ϩ ratio was significantly decreased in the ⌬YHM2 strain (by 58%), more in the ⌬ZWF1 strain (by 87%) and further still in the double knock-out strain (by 91%) as compared with the wild-type strain (Fig. 6B). Again, the NADPH/NADP ϩ ratio was not decreased in ⌬CTP1 cells. Moreover, the changes in this ratio observed in H 2 O 2 -treated ⌬YHM2, ⌬ZWF1, and ⌬YHM2⌬ZWF1 cells were not accompanied by variations in the total amount of NADP ϩ plus NADPH (Fig. 6B). Of note, the total amount of NADP(H) was about three times higher in the presence of H 2 O 2 than in its absence, likely due to the activity of cytosolic Utr1p and Yef1p, which phosphorylate NADH to NADPH. In addition, it is worthwhile mentioning that the NADPH/ NADP ϩ ratio was decreased in the double mutant ⌬YHM2⌬CTP1, with and without H 2 O 2 , to an extent similar to that observed in the single mutant ⌬YHM2 (data not shown). These results indicate that the reduction of the NADPH/NADP ϩ ratio is caused by the absence of YHM2, and not of CTP1. In the next step, the cytosolic levels of NADH and NAD ϩ were measured. In the absence of H 2 O 2 , no significant changes in either the NADH/NAD ϩ ratio or in the total amount of NAD ϩ plus NADH were observed among the different strains (supplemental Fig. S3A). After treatment with H 2 O 2 , the NADH/NAD ϩ ratio (but not the total amount of NAD ϩ plus NADH) decreased in the ⌬YHM2, ⌬ZWF1, and ⌬YHM2⌬ZWF1 strains (supplemental Fig. S3B), although to a lesser extent than the NADPH/NADP ϩ ratio, as compared with wild type. Moreover, the total amount of NADH ϩ NAD ϩ was increased in the presence of H 2 O 2 . The NADH/NAD ϩ ratio was also not affected by H 2 O 2 in the ⌬CTP1 strain as compared with wild type (supplemental Fig. S3B).
Given that NADPH can be oxidized by glutathione reductase, reduced and oxidized glutathione levels in the cytosol were measured. In the absence of H 2 O 2 , the GSH/GSSG ratio was reduced in ⌬YHM2 (by 43%), ⌬ZWF1 (by 55%), and in ⌬YHM2⌬ZWF1 (by 60%) strains but not in ⌬CTP1 cells as compared with wild type (Fig. 7A). In contrast, the total amount of glutathione did not show significant changes among strains. In the presence of H 2 O 2 , the GSH/GSSG ratio was reduced by 35% in ⌬YHM2 cells as compared with wild type without variation in the total amount of glutathione (Fig. 7B). In ⌬ZWF1 cells, the ratio was markedly decreased (by 85%) along with a decrease in total glutathione (by 52%), probably because part of GSH was used to protect proteins from oxidative damage by derivatizing key thiol groups. As compared with ⌬ZWF1 cells, ⌬YHM2⌬ZWF1 cells exhibited a smaller reduction of the GSH/ GSSG ratio (by 72%) together with a smaller decrease in total glutathione (by 30%) with respect to wild-type cells. This finding may be due to the severe stress condition caused by adding H 2 O 2 to the double knock-out, which led to ex novo synthesis of GSH besides protein glutathionylation. In fact, the total amount of glutathione in all the strains was greatly increased in FIGURE 6. NADPH/NADP ؉ ratios in the cytosol of parental and deleted S. cerevisiae strains. Yeast strains, precultured in YP with glucose and grown in acetate-supplemented SC, were incubated in the presence (B) or absence (A) of H 2 O 2 for 60 min. NADPH/NADP ϩ ratios and the sums of NADPH ϩ NADP ϩ are reported as means Ϯ S.E. of at least three independent experiments. Differences between the NADPH/NADP ϩ ratio of ⌬YHM2, ⌬ZWF1, and ⌬YHM2⌬ZWF1 cells and control (wild-type (WT) cells), with and without H 2 O 2 , were significant (*, p Ͻ 0.05, and **, p Ͻ 0.01, one-way ANOVA followed by Bonferroni's t test).
the presence of H 2 O 2 as compared with its absence. Again, the ratio between GSH and GSSG was not significantly decreased in ⌬CTP1 cells (Fig. 7B).
Nicotinamide Coenzyme Levels in Mitochondria-Once ascertained that deletion of YHM2, ZWF1, or both decreases the cytosolic NADPH/NADP ϩ ratio, especially upon H 2 O 2 treatment, this ratio was determined in the mitochondria of knock-out strains. With no addition of H 2 O 2 , the mitochondrial NADPH/NADP ϩ ratio was approximately the same in all the evaluated strains (Fig. 8A). However, upon addition of H 2 O 2 , this ratio was significantly higher in ⌬YHM2 and ⌬YHM2⌬ZWF1 strains as compared with the wild-type strain (Fig.  8B). In all strains, the total amount of NADPH and NADP ϩ coenzymes was similar regardless of H 2 O 2 treatment. Therefore, under oxidative stress, the NADPH/NADP ϩ ratio of ⌬YHM2 and ⌬YHM2⌬ZWF1 cells diminished in the cytosol (Fig.  6B) and increased in mitochondria (Fig. 8B), suggesting that in wildtype cells reducing equivalents of NADPH can be transferred from the mitochondria to the cytosol through the activity of Yhm2p. In untreated wild-type and knock-out cell mitochondria, the NADH level was too low, as observed previously in wild-type cells (15), to assess the NADH/NAD ϩ ratio. Upon addition of H 2 O 2 , both the low NADH/ NAD ϩ ratio and total amount of NADH ϩ NAD ϩ were not significantly different among the strains tested (supplemental Fig. S4B).
Cytosolic Citrate and Oxoglutarate Levels-Cytosolic samples of ⌬YHM2, ⌬ZWF1, and ⌬YHM2⌬ZWF1 cells were also analyzed for the two exchanging metabolites citrate and oxoglutarate. Of note, the cytosol of H 2 O 2treated wild-type cells contained much more citrate than the cytosol of untreated cells (217 nmol/mg protein as compared with 37 nmol/mg protein without H 2 O 2 ; Fig. 9, A and B), presumably due to the activity of the extra-mitochondrial citrate-producing enzyme Cit2p. Moreover, the cytosol of H 2 O 2 -treated ⌬YHM2, ⌬ZWF1, and ⌬YHM2⌬ZWF1 cells contained a significantly lower amount of citrate (by about 27%) than the same fraction of wild-type cells (Fig. 9B). On the contrary, with no treatment the cytosolic level of citrate was approximately the same in the deleted strains as the in wild-type cells (Fig. 9A). Concerning oxoglutarate, without H 2 O 2 treatment, its level in the cytosol was similar in ⌬ZWF1, higher in ⌬YHM2, and significantly higher in ⌬YHM2⌬ZWF1 cells as compared with wild-type cells FIGURE 7. GSH/GSSG ratios in the cytosol of parental and deleted S. cerevisiae strains. Yeast strains, precultured in YP with glucose and grown in acetate-supplemented SC, were incubated in the presence (B) or absence (A) of H 2 O 2 for 60 min. GSH/GSSG ratios and the sums of GSH ϩ GSSG are reported as means Ϯ S.E. of at least three independent experiments. Differences between the GSH/GSSG ratio of ⌬YHM2, ⌬ZWF1, and ⌬YHM2⌬ZWF1 cells and control (wild-type (WT) cells), with and without H 2 O 2 , were significant (*, p Ͻ 0.05, and **, p Ͻ 0.01, one-way ANOVA followed by Bonferroni's t test). Differences between the sum of GSH ϩ GSSG of ⌬ZWF1 and ⌬YHM2⌬ZWF1 and control (wild-type cells) in the presence of H 2 O 2 were significant (*, p Ͻ 0.05, one-way ANOVA followed by Bonferroni's t test). FIGURE 8. NADPH/NADP ؉ ratios in the mitochondria of parental and deleted S. cerevisiae strains. Yeast strains, precultured in YP with glucose and grown in acetate-supplemented SC, were incubated in the presence (B) or absence (A) of H 2 O 2 for 60 min. NADPH/NADP ϩ ratios and the sums of NADPH ϩ NADP ϩ are reported as means Ϯ S.E. of at least three independent experiments. Differences between the NADPH/NADP ϩ ratio of ⌬YHM2 and ⌬YHM2⌬ZWF1 cells and control (wild-type (WT) cells) in the presence of H 2 O 2 were significant (*, p Ͻ 0.05, one-way ANOVA followed by Bonferroni's t test). (Fig. 9A). By contrast, in the presence of H 2 O 2 the cytosolic level of oxoglutarate was moderately (but significantly) reduced in ⌬YHM2 and more so in ⌬ZWF1 and ⌬YHM2⌬ZWF1 cells as compared with wild-type cells (Fig. 9B). This indicates that under these stress conditions oxoglutarate is most likely used to produce the antioxidant tripeptide GSH.
Antioxidant Response to H 2 O 2 Exposure-The intracellular DCF fluorescence (i.e. level of intracellular non-scavengered ROS (18)) of ⌬YHM2 cells exposed to H 2 O 2 was slightly higher than that of H 2 O 2 -treated wild-type cells (Fig. 10). In contrast, the fluorescence increased to 131 Ϯ 7 and 173 Ϯ 5% in the ⌬ZWF1 and ⌬YHM2⌬ZWF1 cells, respectively, after a 90-min H 2 O 2 exposure compared with that of treated wild-type cells (Fig. 10). These findings indicate that the cells lacking both YHM2 and ZWF1 responded to oxidative insult much less efficiently than the individually deleted strains. The ROS scavenger ability of wild-type and deleted cells was further assessed by cytofluorimetry. The histograms of time-dependent fluorescence-activated cell sorter analysis (Fig. 11) show that the intensity of cell fluorescence was markedly increased after a 30

DISCUSSION
In this work, overexpression in S. cerevisiae of a previously unidentified MC and reconstitution of the recombinant protein in liposomes as well as phenotype analysis of yeast knock-out cells have been employed to investigate the function of Yhm2p. The transport properties and kinetic parameters of Yhm2p, together with its mitochondrial localization, demonstrate that this protein is a novel mitochondrial citrate-oxoglutarate carrier. Besides transporting these substrates with great efficiency, Yhm2p also translocates oxaloacetate, fumarate, succinate, and to a much lesser extent, L-malate and other tricarboxylates.
As we have found an accumulation of citrate in ⌬YHM2 cell mitochondria, Yhm2p appears to be essential for the export of citrate from the mitochondria to the cytosol, at least under the experimental conditions employed in this study. In the cytosol, citrate can be used, via isocitrate, to produce oxoglutarate and NADPH by the action of the NADP ϩ -dependent cytosolic isocitrate dehydrogenase, Idp2p (Fig. 12). Therefore, a primary physiological function of Yhm2p is to increase cytosolic NADPH, which is required for various biosynthetic reactions and thiol-dependent peroxidases that utilize reduced glutathione or thioredoxin to protect the cell from oxidative damage. Moreover, because Yhm2p functions exclusively by antiport and oxoglutarate is efficiently transported by this carrier, oxoglutarate can enter the mitochondria in exchange for citrate via Yhm2p. In mitochondria, oxoglutarate can be converted to isocitrate by the NADP ϩ -dependent mitochondrial isocitrate dehydrogenase, Idp1p, and then to citrate, leading to net export of NADPH-reducing equivalents (Fig. 12). Thus, Yhm2p may act as a key component of the citrate-oxoglutarate NADPH redox shuttle between mitochondria and cytosol.