The dicarboxylate carrier (DIC) is an integral membrane protein that catalyses a dicarboxylate–phosphate exchange across the inner mitochondrial membrane. We generated a yeast mutant lacking the gene for the DIC. The deletion mutant failed to grow on acetate or ethanol as sole carbon source but was viable on glucose, galactose, pyruvate, lactate and glycerol. The growth on ethanol or acetate was largely restored by the addition of low concentrations of aspartate, glutamate, fumarate, citrate, oxoglutarate, oxaloacetate and glucose, but not of succinate, leucine and lysine. The expression of the DIC gene in wild-type yeast was repressed in media containing ethanol or acetate with or without glycerol. These results indicate that the primary function of DIC is to transport cytoplasmic dicarboxylates into the mitochondrial matrix rather than to direct carbon flux to gluconeogenesis by exporting malate from the mitochondria. The ΔDIC mutant may serve as a convenient host for overexpression of DIC and for the demonstration of its correct targeting and assembly.
Although the transport of certain metabolites across the mitochondrial inner membrane has been extensively investigated, knowledge about the role of individual mitochondrial transporters in cellular and mitochondrial metabolism is still limited. To gain insight on the physiological role of the DIC we have deleted the nuclear gene encoding DIC in S. cerevisiae. In this paper the effects of the deletion on the metabolism were studied in mitochondria and whole cells. ΔDIC cells did not grow on ethanol or acetate. The addition of substrates that provide Krebs cycle intermediates restored cell viability. We propose that net uptake of dicarboxylates into mitochondria is the primary function of the DIC in yeast. This function is discussed in connection to the role played by the newly discovered succinate–fumarate carrier (SFC) in gluconeogenesis (Palmieri et al., 1997).
Mitochondria lacking DIC are impaired in malate–phosphate exchange
The nuclear gene encoding the mitochondrial dicarboxylate carrier of S. cerevisiae was deleted by homologous recombination of the HIS3 gene. The deletion mutant (ΔDIC strain) was able to grow on 3% glycerol at the same rate and to the same extent as the wild-type yeast. In a first set of experimental conditions we tested the S. cerevisiaeΔDIC strain, the parental strain and the deletion strain transformed with the DIC-pRS416 construct (DIC-pRS416 strain) for the presence of DIC in the mitochondria with an anti-DIC polyclonal antiserum raised against purified DIC in a rabbit. With the deletion mutant no immunodecoration was observed (Fig. 1). In contrast, the wild-type and the DIC-pRS416 mitochondria showed a single immunoreactive band with an apparent molecular mass of about 31.5 kDa, a value that is in good agreement with the mass calculated from the primary sequence of DIC (32991, including the initiator methionine residue) (Palmieri et al., 1996). The amount of DIC in mitochondria from the ΔDIC strain harbouring the plasmid DIC-pRS416 was about threefold higher than in mitochondria from the wild-type strain, as revealed by quantitative Western blotting (data not shown). To exclude the possibility of a disturbance of the overall protein content by the mutation, we immunodecorated wild-type, ΔDIC and DIC-pRS416 mitochondria with polyclonal antisera directed against different mitochondrial carriers, the succinate–fumarate carrier (SFC), the ADP/ATP carrier (AAC) and the phosphate carrier (PiC). In all types of mitochondria, the three mitochondrial carriers were present in practically equal amounts (Fig. 1).
The exchange of the dicarboxylate malate or succinate with phosphate across the mitochondrial inner membrane (which is the defining reaction of the mitochondrial dicarboxylate carrier) can be measured by monitoring the swelling of mitochondria in the presence of an isoosmotic solution (120 mM) of ammonium malate or ammonium succinate plus a catalytic amount of phosphate (Chappel, 1968; Kolarov et al., 1972; Crompton et al., 1974). The results shown in Fig. 2 demonstrate that on addition of phosphate wild-type and DIC-pRS416 mitochondria swelled in either malate or succinate ammonium solution. In contrast to this, ΔDIC mitochondria did not swell in ammonium malate or succinate plus phosphate. As a control, Fig. 2 shows negligible swelling in the ammonium salt of fumarate that is not transported by DIC (Kolarov et al., 1972; Palmieri et al., 1996). Furthermore, all types of mitochondria including those isolated from the ΔDIC strain showed rapid swelling in 120 mM ammonium phosphate, indicating the presence of an active PiC and the integrity of the mitochondria (Fig. 2). Taken together these results demonstrate that no other mitochondrial transport system is redundant with the DIC and that the absence of DIC does not affect the expression and the activity of other carriers.
In the next step, we assessed the membrane potential (Δψ) of ΔDIC mitochondria using the fluorescence dye DiSC3(5) (Sims et al., 1974). This was done by measuring the difference between the fluorescence after addition of mitochondria and substrates, and that after the subsequent addition of valinomycin (in the presence of K+), which represents a relative assessment of Δψ (Zara et al., 1996; Dekker et al., 1997). The difference in fluorescence was virtually the same in ΔDIC as in wild-type and DIC-pRS416 mitochondria (data not shown). We concluded from these results that the DIC gene deletion did not influence the coupling of the inner mitochondrial membrane.
Functional high-level expression of DIC in the deletion strain
We used the DIC deletion strain and the DIC-pYES2 plasmid to obtain overexpression of the DIC protein in yeast mitochondria. SDS–PAGE and Western blotting of mitochondria isolated from the plasmid DIC-pYES2-complemented ΔDIC strain (DIC-pYES2 strain) demonstrated the high-level expression and targeting of the episomal DIC to mitochondria (Fig. 3A and B, lanes 2). The overexpressed band of 31.5 kDa, i.e. the carrier protein, accounted for about 18% of the total mitochondrial protein as estimated by scanning densitometric analysis. Using quantitative Western blotting, the abundance of DIC was calculated to be 2.2 ± 0.4 mg l−1 culture of DIC-pYES2 yeast strain and 60 ± 12 pmol mg−1 mitochondrial protein in the wild-type strain (means ± SD of four determinations). Virtually all mitochondrial carrier proteins have been purified using hydroxyapatite as the major chromatographic step (for references see Palmieri, 1994), because they are not bound to hydroxyapatite in their native conformation (Klingenberg and Lin, 1986) in contrast to most other mitochondrial proteins. To test whether the overexpressed DIC retains the property of the native molecule to pass-through hydroxyapatite, which also reflects the correct assembly in the mitochondrial membrane (Schleyer and Neupert, 1984; Zara et al., 1991), we subjected the Triton X-114-solubilized mitochondria derived from the DIC-pYES2 strain to hydroxyapatite chromatography. As illustrated in 3Fig. 3C, lane 2, only three protein bands were present in the hydroxyapatite pass-through. Out of these bands, the one in the middle of 31.5 kDa was identified as the DIC by Western blot analysis (Fig. 3D).
The high activity of the overexpressed DIC was demonstrated by the rapid swelling of the DIC-pYES2 mitochondria in ammonium malate (or succinate) plus phosphate (Fig. 2). The same mitochondria swelled in ammonium phosphate but not in ammonium salts of impermeable anions such as fumarate or chloride (Fig. 2). Furthermore, they exhibited a Δψ after addition of substrates similar to that observed with wild-type mitochondria (data not shown). Finally, no degradation of overexpressed DIC was revealed by the immunoblot analysis of 3Fig. 3B, lane 2. These results show that the high-level expression of DIC does not disturb the integrity and the stability of the mitochondria. In other experiments the activity of the overexpressed DIC was measured as 0.1 mM [14C]-malate uptake by proteoliposomes, that were reconstituted with Triton X-114 extract of DIC-pYES2 mitochondria and contained 20 mM phosphate (malate–phosphate exchange). The results are shown in Table 1 and demonstrate that the proteoliposomes catalysed a very active uptake of [14C]-malate in exchange for intraliposomal phosphate. Small but significant DIC activity was also observed upon reconstitution of the extract from DIC-pRS416 mitochondria. In contrast, the DIC activity of the ΔDIC mitochondrial extract was not detectable. The reconstituted [14C]-malate–phosphate exchange was markedly inhibited by butylmalonate, the SH-blocking reagent mersalyl, and by the DIC substrates malate, malonate, succinate, phosphate and sulphate (Table 1). In contrast, fumarate, citrate, ADP and glutamate had virtually no effect on the exchange. This means that the overexpressed DIC shows the same specificity as that observed for the native DIC and for the DIC expressed in E. coli (Bisaccia et al., 1988; Lancar-Benba et al., 1996; Palmieri et al., 1996). The basic kinetic data of the overexpressed reconstituted DIC were determined by measuring the initial transport rate at various external malate concentrations in the presence of a constant saturating concentration (20 mM) of internal phosphate. A Km of 0.61 ± 0.09 mM for malate uptake and a Vmax of 5.2 ± 1.4 mmol min−1 g−1 protein were obtained in four experiments (not shown). These kinetic parameters closely resemble the values reported for the yeast DIC (Palmieri et al., 1996). Thus, the overexpressed yeast DIC is functionally active and exhibits all the characteristic properties of the native DIC.
Table 1. . Activity of DIC in liposomes reconstituted with mitochondrial yeast extracts from the ΔDIC strain, the parental strain and the deletion strain transformed with either the DIC-pRS416 or the DIC-pYES2 plasmid. The proteoliposomes were preloaded with 20 mM phosphate and transport was started by adding 0.1 mM [14C]-malate. External anions and inhibitors were added together with [14C]-malate at a concentration of 2 mM, except mersalyl, which was added 2 min before the labelled substrate at a concentration of 0.1 mM.
ΔDIC yeast cells are not able to grow on ethanol or acetate as the sole carbon source
The ΔDIC strain was tested for its ability to utilize different carbon sources (see Table 2). This strain did not grow on synthetic minimal medium (SM) containing 2% ethanol or 3% acetate, as the sole carbon source, even after 1 week of incubation. However, it exhibited substantial growth on all the other non-fermentative (3% glycerol, 2% lactate, 2% pyruvate and 10 mM oxaloacetate) and fermentative (2% glucose and 2% galactose) carbon sources tested, similarly to the wild-type strain. Growth on SM supplemented with either ethanol or acetate was fully restored by complementing the deletion strain with the DIC-pRS416 plasmid (not shown), demonstrating that this difference in phenotype was the result of the absence of DIC protein and not a secondary effect. Interestingly, the growth of the deletion strain on acetate or ethanol was largely restored by the addition of 0.5–1.0 mM aspartate, glutamate, fumarate, citrate and α-oxoglutarate or 0.1 mM oxaloacetate and glucose (see Table 2). It has to be noted that these supplements, when added at the same concentrations alone (i.e. without acetate or ethanol) did not sustain the growth of both the ΔDIC and the parental strains (not shown). As reported for aspartate and oxaloacetate in Table 2, at concentrations lower than those indicated above, the extent of the DIC strain growth on ethanol or acetate is dependent on the supplement concentration. In contrast, 1.0 mM succinate or 1.0 mM leucine and lysine, i.e. amino acids that do not generate Krebs cycle intermediates, did not restore the growth of the ΔDIC strain in the presence of ethanol or acetate.
Table 2. . Growth of WT and ΔDIC yeast under various culture conditions. The values of optical density at 600 nm given in the table refer to cell cultures after 96 h of growth on SM containing the indicated carbon sources and supplements. When necessary, i.e. above an optical density of 0.7, the measurements were made on samples diluted 1:5.
Regulation of the yeast DIC gene expression
As demonstrated above, the DIC activity is essential for growth on ethanol or acetate. Therefore, we decided to compare the expression of DIC in the wild-type strain after growth on glycerol (3%), ethanol (2%), acetate (3%), glycerol plus ethanol or glycerol plus acetate. The immunodecoration of mitochondria isolated from wild-type cells grown on the above-mentioned carbon sources (Fig. 4A and B) revealed that the expression of the DIC gene is repressed by ethanol and by acetate when added both alone or together with glycerol. Unlike DIC, the other three mitochondrial carrier proteins tested (SFC, AAC and PiC) were present in similar amounts under all growth conditions used (Fig. 4).
Assembly of endogenous and overexpressed DIC in mitochondrial inner membrane
We knocked out the DIC gene in S. cerevisiae and found that the absence of DIC activity did not affect cellular respiration, the expression of other mitochondrial carriers and the stability of the mitochondria. These parameters were also unaffected by reintroduction of the DIC gene in the ΔDIC background on either a low- or high-copy episomal vector. The results reported in this paper indicate that overexpressed active DIC is correctly targeted and assembled in yeast, which is an interesting advantage over bacterial expression where DIC accumulates in the cytoplasm as inclusion bodies and has to be renatured. We think therefore that the overexpression of DIC in yeast that we have described here provides a suitable system for studying structure–function relationships of this transporter for both the targeting of the protein and the transport mechanism.
Saccharomyces cerevisiae can grow aerobically on ethanol as sole carbon source by converting it into acetate via acetaldehyde, and then by activating the acetate to acetyl-CoA, which can be fed into both the glyoxylate and tricarboxylate cycles. The glyoxylate pathway is essential for the utilization of C-2 compounds (Fernandez et al., 1992). Succinate, one of the principal products of the glyoxylate pathway, is produced in the cytosol from isocitrate (Taylor et al., 1996). As succinate dehydrogenase is present only in the mitochondrial matrix, the succinate produced in the cytosol has to be imported into mitochondria. In the yeast S. cerevisiae, we have recently identified two mitochondrial transport systems for succinate: the DIC (Palmieri et al., 1996) and the succinate–fumarate carrier (SFC) (Palmieri et al., 1997). The latter exchanges external succinate for internal fumarate across the mitochondrial inner membrane. Fumarate in the cytoplasm is converted first to malate and then to oxaloacetate (see Fig. 6). The cytoplasmic oxaloacetate is funnelled into the gluconeogenic pathway which is indispensable for S. cerevisiae growth on ethanol or acetate (Gancedo and Serrano, 1989; Fernandez et al., 1994). Although the SFC protein requires fumarate as counter-substrate, DIC catalyses the import of succinate into the mitochondria in exchange for internal phosphate. As the latter is recycled into mitochondria by the phosphate carrier, the combined activity of DIC and phosphate carrier leads to a net uptake of succinate (see Fig. 6). In the wild type the conversion of succinate to fumarate and oxaloacetate within the mitochondrion allows the oxidation of acetyl-CoA produced from ethanol or acetate and triggers the activity of the succinate–fumarate exchange. The role of DIC in providing intermediates of the Krebs cycle in the matrix of the mitochondria agrees with the lack of growth of S. cerevisiaeΔDIC strain on ethanol or acetate (but not on other non-fermentative substrates) and is supported by the effect of aspartate, glutamate, fumarate, citrate, oxoglutarate, glucose and oxaloacetate in restoring the deletion strain viability on ethanol or acetate. We think that the function of these substrates is to generate intramitochondrial oxaloacetate (or other Krebs cycle intermediates) to start the operation of the tricarboxylate cycle. This interpretation is supported by the inability of succinate, leucine and lysine to restore the growth of ΔDIC cells on ethanol or acetate and by the very low concentrations of oxaloacetate required to restore the viability on ethanol or acetate. One could argue that the oxaloacetate, which is produced from fumarate exported to the cytosol by the succinate–fumarate carrier, should be sufficient to activate the Krebs cycle activity. There are two possibilities to explain the lack of growth of ΔDIC mutant on ethanol or acetate as sole carbon source. The first possibility is that succinate–fumarate carrier is not functioning under these conditions because of the absence of fumarate in the mitochondrial matrix, which means that there is no gluconeogenesis nor Krebs cycle activity. The second possibility is that oxaloacetate concentration outside the mitochondria is not high enough to be transported into mitochondria (gluconeogenesis is possible but no Krebs cycle activity). It is noteworthy that we have recently identified an oxaloacetate carrier in yeast, which is different from SFC and DIC, and has a rather low affinity for oxaloacetate (L. Palmieri and V. De Marco, unpublished).
The results reported in this paper show that DIC is not required for gluconeogenesis from pyruvate, lactate and glycerol and question the proposal that this carrier is involved in gluconeogenesis from ethanol or acetate by exporting malate from mitochondria (Wills et al., 1986). An argument against the export of malate is the ΔpH dependence of malate transport across the mitochondrial membrane (Palmieri et al., 1970). Furthermore, a dicarboxylate–dicarboxylate exchange via DIC is unlikely because of the high intramitochondrial phosphate concentration (Siess et al., 1982). Finally, after the recent discovery of SFC in yeast (Palmieri et al., 1997) the proposed role of DIC for gluconeogenesis is unnecessary. In the presence of ethanol or acetate, when sufficient Krebs cycle intermediates are provided by DIC, it is useful for S. cerevisiae cells to direct the carbon flux mainly to synthetic pathways. In line with this, one can understand one of our main findings, i.e. the repression of DIC expression by ethanol or acetate. Clearly, a lower activity of DIC would favour the utilization of cytosolic succinate by SFC and thereby the gluconeogenetic pathway. Interestingly, DeRisi et al. (1997) found that among the S. cerevisiae genes encoding mitochondrial carriers only the SFC gene shares a common upstream activating sequence and a similar increased temporal pattern of expression (during the diauxic shift from fermentation to respiration) with two key enzymes of the gluconeogenic pathway. Our data also question the proposed role of DIC in gluconeogenesis in mammals (for references see Meijer and Van Dam, 1974). We have recently detected 35 putative members of the mitochondrial carrier family in the yeast genome (Palmieri et al., 1996) and have identified one of this member as an SFC quite unexpectedly (Palmieri et al., 1997), because it had never been suggested before from experiments in intact mitochondria. In consideration of these findings, it is likely that even more than 35 mitochondrial carriers are present in mammalian mitochondria. At present the existence of other transport systems in this organelle that are involved in directing carbon flux to gluconeogenesis cannot be excluded.
Yeast strains, expression plasmids and synthetic media
The deletion of the DIC gene was accomplished by homologous recombination of the auxotrophic marker HIS3 at the DIC gene locus of S. cerevisiae YPH499 strain. The genotype is MATa ade2-101 his3-Δ200 leu2-Δ1 ura3-52 trp1-Δ63 lys2-801 DIC ::HIS3. The plasmids DIC-pRS416 and DIC-pYES2 were constructed by cloning the DIC ORF into the vector pRS416 (Sikorski and Hieter, 1989) and the vector pYES2 (Invitrogen) behind the inducible Gal–Cyc promoter respectively. These vectors were introduced into the ΔDIC strain. Transformants were selected for uracil auxotrophy. Yeasts were precultured in synthetic complete medium (SC) lacking uracil in the plasmid-complemented strains and supplemented with 3% glycerol. The precultures were diluted with synthetic minimal medium (SM) until a final optical density of 5 × 10−3 at 600 nm was reached. SM contained 0.67% yeast nitrogen base without amino acids, 0.1% KH2PO4, 0.12% (NH4)2SO4, 200 mg l−1 leucine, 40 mg l−1 lysine, 40 mg l−1 tryptophan, 40 mg l−1 uracil, 10 mg l−1 adenine and, in addition, one of the carbon sources indicated in Table 2. The final pH was adjusted to 4.5 or, with acetate or pyruvate to 6.5.
Preparation of yeast mitochondria
For the preparation of mitochondria yeast cells were grown as previously described (Dekker et al., 1997). Only in the case of the DIC-pYES2 strain a second preculture was made with 3% glycerol and 0.1% glucose as carbon sources, and 0.4% galactose was added 6 h before harvesting. Mitochondria were isolated according to standard procedures (Daum et al., 1982).
Swelling of yeast mitochondria
The rate of mitochondrial swelling was monitored by recording the decrease in A546 as previously described (Zara et al., 1996). Yeast mitochondria (100 μg of protein) were added to a glass cuvette containing 1 ml of 120 mM ammonium salts of the indicated anions, 20 mM Tris, 1 mM EDTA, 5 μM rotenone and 0.1 μM antimycin, pH 7.4.
Reconstitution of DIC into liposomes and transport measurements
Mitochondria (200 μg protein ml−1) were solubilized in 3% Triton X-114, 50 mM NaCl, 1 mM EDTA and 10 mM Pipes (pH 7). After 20 min at 0°C, the mixture was spun at 13 8000 × g for 30 min. The supernatant (extract) was reconstituted into liposomes by cyclic removal of the detergent with a hydrophobic column (Palmieri et al., 1995). The mixture used for reconstitution consisted of 30 μl of extract (about 5 μg of protein), 70 μl of 10% Triton X-114, 8.5 mg of egg yolk phospholipids (Fluka) in the form of liposomes, 0.8 mg cardiolipin (Sigma), 20 mM phosphate and 10 mM Pipes (pH 7.0) in a final volume of 700 μl. The mixture was recycled 15 times through an Amberlite XAD-2 (Supelco) column (2 cm × 0.5 cm internal diameter) pre-equilibrated with 10 mM Pipes (pH 7.0) and 20 mM phosphate. After removal of the external substrate by Sephadex G75 chromatography, transport was started by adding [14C]-malate (Amersham) to proteoliposomes and stopped by addition of either 30 mM butylmalonate or a mixture containing 30 mM pyridoxal 5′-phosphate and 10 mM batophenanthroline (Palmieri et al., 1996). In control samples, the inhibitors were added at the beginning together with the labelled substrate. Finally, the external substrate was removed and the radioactivity in the liposomes was measured (Palmieri et al., 1995). For kinetic measurements, the initial transport rate was calculated in mmol min−1 g−1 of protein from the time course of isotope equilibration, as has been published previously (Palmieri et al., 1995).
SDS–PAGE, Western blotting and quantitative immunoblotting were performed as previously described (Zara et al., 1996; Fiermonte et al., 1998). The blue native gel electrophoresis was carried out essentially as described by Palmisano et al. (1998). The membrane potential of mitochondria was assessed by recording the fluorescence changes of the voltage-sensitive dye DiSC3(5) exactly as described in Zara et al. (1996). A partial purification of DIC from DIC-pYES2 mitochondria was achieved by applying 200 μl of mitochondrial extract (30–40 μg protein) onto a hydroxyapatite column (Pasteur pipette containing 0.2 g of cold dry material) and eluting with the solubilization buffer (see above). The first 200 μl of the eluate were collected.
This work was supported by the ‘C.N.R. Target Project on Biotechnology’ and by the Ministero dell'Università e della Ricerca Scientifica e Tecnologica (MURST).