Both authors contributed equally to this work. Dedicated to the memory of Professor Eraldo Antonini.
Identification and metabolic role of the mitochondrial aspartate-glutamate transporter in Saccharomyces cerevisiae
Article first published online: 7 NOV 2003
Volume 50, Issue 4, pages 1257–1269, November 2003
How to Cite
Cavero, S., Vozza, A., Del Arco, A., Palmieri, L., Villa, A., Blanco, E., Runswick, M. J., Walker, J. E., Cerdán, S., Palmieri, F. and Satrústegui, J. (2003), Identification and metabolic role of the mitochondrial aspartate-glutamate transporter in Saccharomyces cerevisiae. Molecular Microbiology, 50: 1257–1269. doi: 10.1046/j.1365-2958.2003.03742.x
- Issue published online: 7 NOV 2003
- Article first published online: 7 NOV 2003
- Accepted 25 July, 2003.
The malate-aspartate NADH shuttle in mammalian cells requires the activity of the mitochondrial aspartate-glutamate carrier (AGC). Recently, we identified in man two AGC isoforms, aralar1 and citrin, which are regulated by calcium on the external face of the inner mitochondrial membrane. We have now identified Agc1p as the yeast counterpart of the human AGC. The corresponding gene was overexpressed in bacteria and yeast mitochondria, and the protein was reconstituted in liposomes where it was identified as an aspartate-glutamate transporter from its transport properties. Furthermore, yeast cells lacking Agc1p were unable to grow on acetate and oleic acid, and had reduced levels of valine, ornithine and citrulline; in contrast they grew on ethanol. Expression of the human AGC isoforms can replace the function of Agc1p. However, unlike its human orthologues, yeast Agc1p catalyses both aspartate-glutamate exchange and substrate uniport activities. We conclude that Agc1p performs two metabolic roles in Saccharomyces cerevisiae. On the one hand, it functions as a uniporter to supply the mitochondria with glutamate for nitrogen metabolism and ornithine synthesis. On the other, the Agc1p, as an aspartate-glutamate exchanger, plays a role within the malate-aspartate NADH shuttle which is critical for the growth of yeast on acetate and fatty acids as carbon sources. These results provide strong evidence of the existence of a malate-aspartate NADH shuttle in yeast.
The transport of metabolites through the inner mitochondrial membrane is accomplished by related proteins belonging to the mitochondrial carrier (MC) superfamily (Palmieri, 2003). MC proteins display three repeated regions, each about 100 amino acids long, which contain a conserved sequence motif and two putative transmembrane domains. Over the past few years, a number of new MCs have been identified [(Fiermonte et al., 2001, 2002; Marobbio et al., 2002; Hoyos et al., 2003) and references therein], including members of a subfamily of calcium-binding mitochondrial carriers (CaMCs) with a characteristic bipartite structure. Their C-terminal domains have the common structural features of the MC superfamily, whereas their N-terminal domains harbour EF-hand calcium-binding motifs (del Arco and Satrústegui, 1998; del Arco et al., 2000). Aralar1 and citrin are two closely related human CaMCs expressed in different tissues (del Arco and Satrústegui, 1998; del Arco et al., 2000; 2002; Begum et al., 2002; Ramos et al., 2003). Mutations in citrin cause type II citrullinaemia (Kobayashi et al., 1999). We have recently found that aralar1 and citrin are isoforms of the mitochondrial aspartate/glutamate carrier (AGC) (Palmieri et al., 2001a).
In mammalian mitochondria, AGC plays a key role in the malate-aspartate NADH shuttle which conveys reducing equivalents from cytosolic NADH into the matrix and finally to the mitochondrial electron transport chain. Until now, the existence of a functional malate-aspartate NADH shuttle in Saccharomyces cerevisiae has been a matter of debate. The presence of external NADH dehydrogenases that oxidize NADH on the external face of the inner mitochondrial membrane (Luttik et al., 1998) together with the absence of a classic complex I coupled to the generation of proton-motive force has raised doubts that NADH shuttles are needed in this organism. However, the operation of as yet unidentified redox shuttles that transfer reducing equivalents across the yeast mitochondrial membrane has been shown (Bakker et al., 2001). The enzymes of the malate-aspartate shuttle are all present in yeast, even though the mitochondrial localization of Aat1p which presumably corresponds to the mitochondrial aspartate aminotransferase (Morin et al., 1992) has not been experimentally verified. The existence in yeast of the two inner membrane components of the malate-aspartate shuttle, i.e. the oxoglutarate carrier (OGC) and the AGC has also been questioned. However, two mitochondrial transporters, Odc1p and Odc2p, have been recently identified in yeast that are able to transport oxoglutarate although they are not orthologues of the mammalian OGC (Palmieri et al., 2001b). They transport 2-oxoadipate, 2-oxoglutarate and malate in reconstituted liposomes by a strict exchange mechanism (Palmieri et al., 2001b). The main function of Odc1p and Odc2p is probably to export 2-oxoadipate and 2-oxoglutarate for lysine and glutamate synthesis (Palmieri et al., 2001b). Recent observations, however, indicate that they may also function as oxoglutarate/malate exchangers (Tibbetts et al., 2002).
Aralar1 and citrin are related to Agc1p, a member of the CaMC subfamily in S. cerevisiae (902 amino acids, open reading frame YPR021c), although its N-terminal domain does not have EF-hand motifs. This prompted us to investigate the function of Agc1p. Previous studies demonstrated that the C-terminal domains (CTD) of aralar1 and citrin fully account for their transport properties and that their activity is stimulated by Ca2+ binding to the N-terminal domains (Palmieri et al., 2001a). The CTD of Agc1p was overexpressed in Escherichia coli and reconstituted into phospholipid vesicles where it transports aspartate and glutamate by both exchange and uniport mechanisms. We show that the yeast AGC operates with the same transport properties in vivo as in vitro. In addition, to gain insight into the physiological role of the Agc1p we deleted the AGC1 gene in S. cerevisiae and studied the resulting knock-out (agc1Δ) cells using conventional metabolic approaches as well as 13C nuclear magnetic resonance (NMR). Our results provide evidence that Agc1p is a yeast mitochondrial transporter for aspartate and glutamate. It plays a role in nitrogen metabolism and ornithine synthesis, and is required for the operation of the malate-aspartate redox shuttle in S. cerevisiae, a NADH shuttle necessary for growth on acetate and fatty acids as carbon sources.
Bacterial expression of Agc1p CTD
The C-terminal domain of Agc1p (Agc1p CTD, corresponding to residues 508–902 of the protein and containing the full MC homologous region; calculated MW 44742) was expressed at high levels in Escherichia coli BL21(DE3) (Fig. 1, lane 4). It accumulated as inclusion bodies, and was purified by centrifugation and washing (Fig. 1, lane 5) with a yield of 10–20 mg l−1 bacterial culture. The protein was not detected in bacteria harvested immediately before induction of expression (Fig. 1, lane 2), nor in cells harvested after induction but lacking the coding sequence in the expression vector (Fig. 1, lane 3). The identity of the expressed protein was confirmed by N-terminal sequencing.
Functional characterization of recombinant Agc1p CTD
The Agc1p CTD was reconstituted into liposomes and its transport properties were tested in homo-exchange (same substrate inside and outside) experiments. Using external and internal substrate concentrations of 1 and 10 mM, respectively, the reconstituted protein catalysed active l-[14C]glutamate/l-glutamate and l-[14C]aspartate/l-aspartate exchanges that were inhibited by a mixture of HgCl2 and pyridoxal 5′-phosphate. No homoexchange activities were measured for phosphate, malate, citrate, carnitine and 2-oxoglutarate (data not shown).
The l-[14C]aspartate/l-aspartate and l-[14C]glutamate/l-glutamate exchange reactions catalysed by reconstituted Agc1p CTD followed first order kinetics (rate constants 0.020 and 0.021 min−1 respectively), isotopic equilibrium being approached exponentially (Fig. 2A and B). In addition, when the proteoliposomes were preloaded with NaCl instead of substrate, a low but significant uptake of l-aspartate and especially of l-glutamate was observed (Fig. 2A and B). With internal NaCl, the initial transport rates were 0.41 ± 0.09 and 0.87 ± 0.22 µmol min−1 g−1 protein for aspartate and glutamate, respectively, in four experiments. Unidirectional transport can be studied more conveniently by efflux measurements of reconstituted proteoliposomes (Palmieri et al., 1995). These assays require removal of the external labelled substrate by exclusion chromatography after equilibration of the radioactivity between the intraliposomal and the external compartments and then incubation of the labelled proteoliposomes in the absence of external substrate. Figure 2C shows that a substantial efflux of aspartate and especially glutamate had occurred already during the chromatographic step, which was accomplished in 15 min. Further efflux of aspartate was observed in the subsequent incubation with buffer alone (Fig. 2D). This efflux was abolished by the presence of the inhibitors HgCl2 and pyridoxal 5′-phosphate and was enhanced by the addition of external aspartate (Fig. 2D). The addition of buffer alone at pH 7.5 induced greater efflux of aspartate than buffer at pH 6.5 (Fig. 2D), suggesting that substrate uniport is dependent on the transmembrane pH gradient.
The substrate specificity of reconstituted Agc1p CTD was examined by measuring the uptake of l-[14C]aspartate and l-[14C]glutamate into proteoliposomes that had been preloaded with various potential substrates. As shown in Table 1, the highest activities were observed in the presence of internal l-aspartate and l-glutamate. Much lower activities were obtained with the d-isomers of aspartate and glutamate. To a lesser extent, l-cysteinesulfinate and l-cysteate were exchanged with external labelled substrates. In addition, l-homocysteinesulfinate and l-homocysteate were also transported by Agc1p as demonstrated by the uptake of l-[14C]glutamate. Very low activities were found with all of the other compounds listed in Table 1 and (not shown) with fumarate, succinate, malate, malonate, citrate and oxaloacetate. The residual activity in the presence of these substrates was virtually the same as the activity observed in the presence of NaCl.
|Internal substrate||Substrate transport (nmol min−1 mg−1 protein)|
|l-[14C] aspartate||l-[14C] glutamate|
|None (Cl– present)||0.15||0.35|
The l-[14C]aspartate/l-aspartate and l-[14C]glutamate/l-glutamate exchange reactions catalysed by Agc1p CTD were inhibited strongly (97–100%) by 50 mM pyridoxal 5′-phosphate and 100 µM mercurials (HgCl2, mersalyl and p-chloromercuriphenylsulfonate). They were also inhibited (78–83%) by 1 mM N-ethylmaleimide and 20 mM diethyl pyrocarbonate, but not by known inhibitors of other mitochondrial carriers [i.e. butylmalonate, phenylsuccinate and 1,2,3-benzenetricarboxylate (2 mM each), 20 µM carboxyatractyloside, 10 mM bathophenanthroline and 0.1 mM α-cyano-4-hydroxycinnamate]. These results show that the inhibition characteristics of the recombinant Agc1p CTD, except for the effect of bathophenanthroline, are similar to those previously described for aralar1 and citrin (Palmieri et al., 2001a).
The influence of the membrane potential was investigated on the l-aspartate/l-glutamate exchange mediated by the recombinant protein Agc1p CTD. With Agc1p CTD the rate of the l-[14C]aspartateout/l-glutamatein exchange was stimulated by a K+ diffusion potential generated across the proteoliposomal membranes with valinomycin/KCl (calculated value = 100 mV, positive inside) (Table 2). In the absence of a preformed K+ gradient or when the l-[14C]aspartateout/l-aspartatein homo-exchange was measured, no effect was observed. Therefore, the data are in agreement with Agc1p catalysing an electrogenic exchange of negatively charged l-aspartate for electroneutral l-glutamate.
|Internal substrate||K+in/K+out (mM/mM)||Aspartate uptake (nmol min−1 mg−1 protein)|
Kinetic characteristics of recombinant Agc1p CTD
The kinetic parameters of Agc1p CTD were investigated by analysing the rate of uptake of l-[14C]aspartate or l-[14C]glutamate into proteoliposomes (at 25°C, pH 6.5) at various external substrate concentrations and a constant internal aspartate concentration (20 mM). The Vmax of the glutamate/aspartate exchange (18.7 ± 1.5 µmol min−1 g−1 protein) was 30–40% lower than that of the aspartate/aspartate exchange (29.4 ± 2.6 µmol min−1 g−1 protein), in agreement with the electrogenic nature of the glutamate/aspartate exchange (LaNoue et al., 1974; Palmieri et al., 2001a). The half-saturation constant (Km) for external glutamate was six times higher than for aspartate (the mean values from four experiments are 0.25 ± 0.04 mM and 40 ± 5 µM for glutamate and aspartate respectively).
Agc1p operates in yeast mitochondria
We used the YPH499 strain and the pYES2-AGC1 plasmid to obtain overexpression of the Agc1p in yeast mitochondria. The activity of the overexpressed Agc1p was measured as l-[14C]aspartate uptake by proteoliposomes that were reconstituted with Triton X-100 extract of pYES2-AGC1 mitochondria and contained 20 mM l-aspartate (l-aspartate/l-aspartate exchange) or l-glutamate (l-glutamate/l-aspartate exchange). The results shown in Fig. 3 demonstrate that these proteoliposomes catalysed an active uptake of L-[14C]aspartate in exchange for intraliposomal substrates. The exchange activities were three to fivefold higher than in proteoliposomes reconstituted with wild-type mitochondrial extracts. In contrast, virtually no l-[14C]aspartate uptake was observed with extracts of Yagc1Δ mitochondria (an AGC1 deletion strain obtained from YPH499). As a control, the phosphate/malate exchange (Fig. 3) and the oxoglutarate/oxoglutarate exchange (not shown) were virtually the same in wild-type, Yagc1Δ and pYES2-AGC1 mitochondrial extract proteoliposomes.
In yeast, the glutamate transported into the mitochondrial matrix is transaminated to α-ketoglutarate that can be decarboxylated. Because CO2 production from l-[1-14C]glutamate reflects the ability of mitochondria to internalize this compound, we analysed the l-[1-14C]glutamate decarboxylation in mitochondria from wild-type and Wagc1Δ (an AGC1 deletion strain obtained from W303) cells. Figure 4A shows that l-[1-14C]glutamate conversion into CO2 by wild-type mitochondria is completely blocked in the presence of 5 mM aminooxyacetate, an inhibitor of aminotransferases. This result is consistent with the lack of glutamate dehydrogenase in yeast mitochondria (Hollenberg et al., 1970). Figure 4B shows that the production of CO2 from l-[1-14C]glutamate in Wagc1Δ mitochondria was very small as compared with wild-type mitochondria. The reduced CO2 formation from glutamate was specific for this amino acid, as CO2 production from [1-14C]α-ketoglutarate that also requires α-ketoglutarate dehydrogenase, but not glutamate transport, was similar in both strains (Fig. 4C). Taken together, these results demonstrate that Agc1p functions in yeast mitochondria.
Growth characteristics of agc1Δ mutants
Having established the transport function of Agc1p by in vitro assays, the effect of deleting its gene in yeast cells was investigated. Growth of Wagc1Δ cells on rich medium with glucose (YPD) was indistinguishable from wild-type cells. Differences in growth appeared when cells were grown on oleic acid (YPO) and non-fermentable carbon sources, especially on minimal medium. Figure 5A shows that Wagc1Δ cells did not grow on oleic acid or acetate, and had a lower growth on lactate, pyruvate and oxaloacetate. The differences in growth on acetate were mainly seen from the second day onwards (Fig. 5C). As the W303 strain has objections to growth on minimal medium supplemented with ethanol, we utilized another deletion strain, Yagc1Δ and tested its ability to grow on this carbon source. As shown in Fig. 5B, Yagc1Δ cells exhibit substantial growth on ethanol, similar to the wild-type strain, and have the same growth deficiency on acetate, lactate, pyruvate and oxaloacetate already described for Wagc1Δ cells. The comparable growth on ethanol for wild-type and Yagc1Δ strains excludes the possibility that the lack of growth on acetate or oleic acid for agc1Δ strains is due to a defect of the tricarboxylate and glyoxylate cycles or gluconeogenesis.
Growth on synthetic minimal medium (SM, see Experimental procedures) supplemented with acetate was restored by complementing the deletion strain with the pYX142-AGC1-CTD plasmid but not with the empty pYX142 vector (Fig. 5C), demonstrating that the difference in phenotype is due to the absence of Agc1p and not to a secondary cause. In addition, the expression of citrin and aralar1 in the knock-out strain restored the yeast's ability to grow fully (citrin, Fig. 5C) and partially (aralar1, not shown) on acetate. The reduced restoring effect of aralar1 may be accounted for by the lower turnover number of this protein compared with citrin (Palmieri et al., 2001a) or alternatively by a lower expression in yeast mitochondria (Fig. 5D).
Amino acid composition and 13C NMR studies of acetate metabolism in agc1Δ yeast
Because Agc1p activity is essential for growth on acetate (Fig. 5), we studied the amino acid composition and the utilization of [2-13C] acetate in the Wagc1Δ and the parental strain (den Hollander et al., 1981; Tran-Dinh et al., 1996). To this end, yeast cells were grown on 2% sodium pyruvate as carbon source, to express maximally genes involved in oxidative and gluconeogenic pathways, and then incubated with [13C] acetate to obtain extracts that were used to study the amino acid composition and acetate utilization.
The amino acid composition of wild-type and Wagc1Δ deletion strains is shown in Table 3. The levels of valine, ornithine, citrulline and leucine in the deletant were markedly lower than in wild-type strain. The levels of other amino acids did not differ significantly between wild-type and Wagc1Δ yeast. Moreover, the decrease in valine, ornithine and citrulline in Wagc1Δ yeast was restored partially in Wagc1Δ(AGC1-CTD) and Wagc1Δ(CITRIN) strains (Fig. 6). Although the leucine level was also restored in the transformants, this result cannot be directly correlated with the expression of AGC1-CTD or CITRIN in the deletion strain because the transformants are not auxotrophic for leucine at variance with the deletion strain.
13C NMR spectra of extracts from yeast cells incubated with [2-13C] acetate during 1–3 h, showed a gradual decrease in the C2 acetate resonance and the appearance of signals corresponding to the C4, C3 and C2 carbon resonances of glutamate, the C4 carbon resonance of glutamine, the C1, C6 and the internal carbon resonances of trehalose (den Hollander et al., 1981). No significant difference was observed between the 13C NMR spectra from wild-type and agc1Δ cells, except that 13C labelling of the internal carbons of trehalose was higher in mutant cells (data not shown). The 13C NMR spectra of extracts from wild-type and mutant cells incubated with 100 mM doubly labelled acetate for 4 h are shown in Fig. 7. All resonances observed with [2-13C] acetate were present in these spectra but were split because of the 13C-13C couplings characteristic of 13C2 acetate incubations (Cerdán et al., 1990). Similar natural abundance of 13C resonances was observed in the spectra obtained from wild-type and agc1Δ yeast incubated with unlabelled acetate. These signals were singlets in all cases and significantly smaller than the multiplets observed with 13C-labelled substrates (data not shown). Major labelled peaks were the C1, C2, C3, C4 of glutamate and the trehalose carbons mentioned above, together with the C2 and C1 doublets of acetate. Resonances from alanine (C2, 51.1 p.p.m., and C3, 16.9 p.p.m.), internal carbons of citrate (47.2 p.p.m.), and glutamine C4 (32.4 p.p.m.) and C5 (178 p.p.m.) were also present. The major difference between the wild-type and the agc1Δ strains was a much reduced labelling of alanine (both C2 and C3 carbon resonances), and a stronger labelling of C1, C6 and the internal carbons of trehalose in the mutant. The finding that trehalose labelling from acetate increases in the Wagc1Δ yeast strain indicates that the flux of acetate along the gluconeogenic pathway is higher in the mutant.
Effect of transaminase inhibitors on growth on oleic acid
Peroxisomal β-oxidation of fatty acids results in the production of large amounts of reducing equivalents (NADH) that are shuttled to mitochondria. If the malate-aspartate NADH shuttle is involved, one would expect an inhibition of growth on rich medium with oleate in the presence of the transaminase inhibitor aminooxyacetate. Figure 8 shows that aminooxyacetate completely inhibited growth of wild-type yeast with oleic acid and acetate as carbon sources, whereas growth on glucose or lactate was only slightly inhibited either in wild-type or Wagc1Δ yeast. These results suggest that the malate-aspartate NADH shuttle is specifically required for growth on acetate and fatty acids.
Agc1p as a mitochondrial carrier of aspartate and glutamate
We have identified Agc1p as the yeast orthologue of the human AGC isoforms citrin and aralar1. The substrate specificity, transport affinities, inhibitor sensitivity and voltage dependence properties of the reconstituted protein are similar to those of the human AGCs. However, the yeast protein differs markedly from the mammalian AGCs in its ability to catalyse the uniport of aspartate and, to a greater extent, of glutamate besides the aspartate/glutamate exchange. In addition, the yeast protein differs from the mammalian orthologues because it is able to transport l-homocysteinesulfinate and l-homocysteate (although rather poorly), it transports l-cysteinesulfinate less efficiently and it is insensitive to bathophenanthroline.
The C-terminal domain of Agc1p is more closely related to the corresponding domains of citrin and aralar1 (54%) than is any other yeast member of the MC family, suggesting that Agc1p and citrin/aralar1 originated from a common ancestor. We speculate that this ancestor duplicated in animals giving rise to AGC (strict aspartate-glutamate exchanger) and the related GC [ΔpH-dependent glutamate uniporter (Fiermonte et al., 2002)], whereas in yeast it might have evolved into Agc1p, which possesses both exchange and uniport activities. It is noteworthy that no yeast GC orthologue has yet been identified; Agc1p is the member of the MC family in yeast that has the highest homology with the mammalian GC isoforms (Fiermonte et al., 2002); and all the other yeast carriers that phylogenetically cluster together with Agc1p have already been identified as dicarboxylate or tricarboxylate transporters (Palmieri et al., 2000a, b).
Role of Agc1p as a transporter of glutamate in mitochondria for transamination reactions and ornithine synthesis
The existence of a transporter for glutamate in yeast mitochondria has been previously inferred on the basis of the location of the enzymes responsible for the synthesis of ornithine (Jauniaux et al., 1978), the transamination of pyruvate to alanine (Maaheimo et al., 2001), and the transamination of the branched-chain α-keto acids to the corresponding amino acids (Eden et al., 1996).
Our results indicate that Agc1p provides glutamate to mitochondria, as a consequence of its ability to catalyse the unidirectional uptake of glutamate. This explains the effect of null mutants on the levels or labelling of amino acids that require a transamination reaction in mitochondria for their synthesis, e.g. the drop in valine levels or the reduced alanine labelling from 13C-acetate. Total alanine levels in the experiments with 13C-acetate were unchanged, possibly because pyruvate was used as the carbon source, and therefore cytosolic transamination to alanine probably compensates for the defects in mitochondrial transamination. The role of mitochondrial glutamate as a carbon and nitrogen source for ornithine synthesis explains the drop in ornithine and citrulline levels in AGC null mutants. Because the yeast strains employed in this investigation are auxotrophic for leucine, it was surprising to observe a strongly reduced concentration of this amino acid in agc1Δ compared with wild-type cells. The simplest explanation for this result is that leucine is utilized in agc1Δ mitochondria as an alternative source of ammonium in place of mitochondrial glutamate, via the mitochondrial branched-chain amino acid aminotransferase (Eden et al., 1996; Kispal et al., 1996).
Our results also suggest that yeast cells possess another mitochondrial carrier involved in glutamate uptake. First, the deficiency in glutamate decarboxylation in mitochondria isolated from Wagc1Δ yeast is not complete. Second, a residual glutamate transport activity is observed in Yagc1Δ mitochondrial extract (between 20% and 28%). However, it appears that this further mitochondrial glutamate carrier does not compensate for the absence of Agc1p even in YP rich medium (Fig. 8).
Role of Agc1p in the malate-aspartate NADH shuttle
The results from this study indicate that Agc1p plays a role in the yeast malate-aspartate NADH shuttle and that this role is critical for growth on acetate and fatty acids. The growth of S. cerevisiae on ethanol or acetate requires the operation of the glyoxylate cycle for the biosynthesis of C4 compounds, resulting in NADH overproduction via the cytosolic malate dehydrogenase (Mdh2p). Therefore, an efficient system for the reoxidation of cytosolic NADH is required under these conditions. The significance of the malate-aspartate shuttle in S. cerevisiae has been previously investigated by deleting MDH2 or NDE1, the genes encoding the cytosolic malate dehydrogenase and the principal external NADH dehydrogenase, respectively (Small and McAlister-Henn, 1998). Mutants lacking either MDH2 or NDE1 grow on ethanol or acetate, albeit much more slowly than the wild-type strain, while an MDH2ΔNDE1Δ double mutant does not grow at all on ethanol or acetate. These findings were interpreted as evidence for the operation of a malate-aspartate shuttle in ethanol grown cultures (Small and McAlister-Henn, 1998). However, the phenotype of this mutant does not provide conclusive evidence for the activity of the malate-aspartate shuttle in S. cerevisiae, as cytosolic malate dehydrogenase is also involved in the glyoxylate cycle (van Roermund et al., 1995).
The lack of growth of agc1Δ cells on acetate, but not on ethanol, resembles the growth phenotype observed with the disruption of the MDH1 gene encoding the mitochondrial malate dehydrogenase, another enzyme of the malate-aspartate shuttle (McAlister-Henn and Thompson, 1987). The most likely explanation for the phenotype of cells lacking either Agc1p or Mdh1p is that growth on acetate as the sole carbon source is strictly dependent on the malate-aspartate shuttle to reduce the cytosolic NADH/NAD ratio efficiently. In contrast, during growth on ethanol, the activation of the ethanol-acetaldehyde shuttle system may compensate for the absence of the malate-aspartate shuttle. A compensative effect of the ethanol-acetaldehyde shuttle has been proposed before (von Jagow and Klingenberg, 1970) to explain the growth on ethanol, but not on acetate, of NDIΔ cells devoid of the internal NADH dehydrogenase. It has to be noted that during growth on ethanol, a further compensative effect could be exerted by the glycerol-3-phosphate shuttle as this shuttle is active only in the presence of reduced substrates such as ethanol, and not with more oxidized substrates such as lactate, pyruvate and probably acetate (Larsson et al., 1998). Further evidence that growth on acetate is strictly dependent on the malate-aspartate shuttle is the increased flow of carbon from acetate into trehalose in agc1Δ cells. Indeed, a disruption of the shuttle in agc1Δ cells is expected to favour the metabolic pathways, such as gluconeogenesis, that reduce the cytosolic NADH/NAD ratio.
Growth on oleic acid has been suggested to be dependent on the operation of a redox shuttle in the inner mitochondrial membrane (Tibbetts et al., 2002) in the light of studies on the two isoforms of the yeast oxoglutarate transporter. At 36°C, odc1Δ odc2Δ cells do not grow on oleic acid and grow on acetate only after a significant lag period (Tibbetts et al., 2002). The finding that Agc1p and Odc1p/Odc2p are required for growth on oleate provides evidence that the malate-aspartate NADH shuttle is critical for transporting reducing equivalents from peroxisomal β-oxidation into the mitochondria. A search for the presence of the conserved ORE (oleate response element) sequence CGGNNNTNA-N9-12-CCG (Karpichev and Small, 1998) in the upstream promoter region of the ODC1 locus revealed the presence of two partial ORE consensus sequences (CGGNNNTNC-N9-12-CCG and CGGNNNTNT-N9-12-CCG) upstream of the translation start site (Tibbetts et al., 2002). Interestingly, a partial ORE consensus sequence (CGGNNNTNA-N9-12-CGG) upstream of the translation start site is also present in the AGC1 locus, suggesting that all three genes are induced by oleic acid. In addition, the milder phenotype of the odc1Δodc2Δ strain has been explained by assuming the presence of a third oxoglutarate carrier which might compensate for the loss of Odc1p and Odc2p (Tibbetts et al., 2002). However, the strict requirement of Agc1p for growth on both oleate and acetate indicates that no other mitochondrial carrier in yeast is capable of catalysing the aspartate-glutamate exchange.
As well as for the malate-aspartate cycle, AGC is essential in mammals for the supply of aspartate to the cytosol, where it is required for gluconeogenesis and the urea cycle (Palmieri, 2003; Williamson et al., 1974). The following evidence indicates that Agc1p is not required for these functions in yeast. First, the normal growth of Yagc1Δ yeast on ethanol shows that Agc1p is unnecessary for gluconeogenesis. Also, yeast pyruvate carboxylase, unlike the mammalian enzyme, is cytosolic, making the transport of aspartate out of mitochondria unnecessary for gluconeogenesis. Second, the reactions of the urea cycle have a different location in yeast and mammalian cells. The argininosuccinate synthetase reaction is cytosolic in both cases, but one of its two substrates, citrulline, is produced in the mitochondria of mammals and in the cytosol of S. cerevisiae (Jauniaux et al., 1978). Third, if Agc1p provided aspartate to argininosuccinate synthetase, a decrease in the levels of the products of this and downstream reactions in Agc1p null cells would be expected. However, arginine does not decrease, indicating that Agc1p is not essential for argininosuccinate synthesis.
In summary, yeast Agc1p fulfils two functions. It is involved in glutamate transport into mitochondria for transaminase reactions and for the biosynthesis of amino acids, and it functions as an aspartate-glutamate exchanger within the malate-aspartate NADH shuttle, that is required for growth on acetate and fatty acids. How these two functions and operating modes of the carrier are regulated in vivo is an interesting avenue for future research. Agc1p has a long extension of about 600 amino acids at the N-terminus of the mitochondrial carrier consensus sequence. The human isoforms of AGC harbour EF-hand Ca2+-binding motifs in their N-terminal extensions (del Arco et al., 2000) providing for Ca2+ regulation of the carrier function (Palmieri et al., 2001a). The long N-terminus of Agc1p does not have EF-hands or other Ca2+-binding motifs. This agrees with the known lack of calcium sensitivity of yeast mitochondria (Carafoli et al., 1970; Manon and Guerin, 1993). An interesting possibility is that the Agc1p N-terminus regulates the operation modes of the carrier, perhaps through interaction with other proteins in the intermembrane space, as has been found for the intermembrane space-facing dehydrogenases (Pahlman et al., 2002).
Bacterial expression and purification of Agc1p CTD
Protein was overexpressed at 37°C in E. coli BL21(DE3) as described before (Fiermonte et al., 1993). Inclusion bodies were purified on a sucrose density gradient (Fiermonte et al., 1993), washed at 4°C with TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.0), then twice with a buffer containing Triton X-114 (3%, w/v), 1 mM EDTA and 10 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES)-KOH pH 7.0, and twice again with TE buffer. Agc1p CTD was solubilized in 1.75% sarkosyl (w/v), and a small residue was removed by centrifugation (258 000 g, 1 h). Protein was separated by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate (SDS–PAGE) in 15% gels and stained with Coomassie blue dye. The yield of purified protein was estimated by laser densitometry of stained samples (Fiermonte et al., 1998).
Reconstitution of Agc1p CTD into liposomes
The purified Agc1p CTD in sarkosyl was reconstituted by cyclic removal of detergent with a hydrophobic column as described previously (Palmieri et al., 1995). The composition of the initial mixture used for reconstitution was 42.5 µl Agc1p CTD (about 60 µg), 70 µl 10% Triton X-114, 100 µl egg-yolk phospholipids (Fluka) in the form of sonicated liposomes, 20 mM l-aspartate or l-glutamate (except where otherwise indicated), 0.86 mg ml−1 cardiolipin and 20 mM PIPES, pH 6.0 or 7.0 (see legends to figures) and water to a final volume of 700 µl. This mixture was recycled 13 times through an Amberlite column (Bio-Rad) (3.5 cm × 0.5 cm).
External substrate was removed from proteoliposomes on a Sephadex G-75 column pre-equilibrated with buffer A (50 mM NaCl, 10 mM PIPES) at pH 6.5 (except where indicated). Transport at 25°C was started by adding 0.05 mM l-[14C]aspartate or 0.2 mM l-[14C]glutamate and terminated by addition of 100 µM mercuric chloride and 50 mM pyridoxal 5′-phosphate [the ‘inhibitor-stop’ method (Palmieri et al., 1995)]. In controls, the inhibitors were added with the labelled substrate. The external radioactivity was removed on Sephadex G-75 and the internal radioactivity was measured. Various other transport activities were also assayed by the inhibitor-stop method. For efflux measurements, proteoliposomes containing 1 mM l-glutamate or 1 mM l-aspartate were labelled with 10 µM l-[14C]glutamate or l-[14C]aspartate by carrier-mediated exchange equilibration (Palmieri et al., 1995). After 60 min, the external radioactivity was removed by passing the proteoliposomes through Sephadex G-75 columns, pre-equilibrated with 50 mM NaCl and 0.1 mM PIPES pH 6.5. Efflux was started by adding unlabelled external substrate or buffer A alone at the indicated pH, and terminated by adding the inhibitors indicated above.
Yeast strains, media and growth conditions
The deletion strain Wagc1Δ was generated from W303 (a/α, ura3-1/ura3-1, trp1-Δ2/trp1-Δ2, leu2-3,112/leu2-3,112, his3-11/his3-11, ade2-1/ade2-1, can1-100/can1-100) using the long-flanking homology method (Wach, 1996). The upstream and downstream regions of the AGC1 coding sequence were amplified by PCR, creating an extension homologous to the kanMX4 module [obtained from pFA6-KanMX4 (Wach, 1996)]. Unless otherwise indicated, genomic DNA of the S. cerevisiae wild-type strain was used as a template for PCRs.
The aralar1 and citrin complementing plasmids pYX142-ARALAR1 and pYX142-CITRIN were generated by cloning the PCR products of human ARALAR 1 and CITRIN coding sequence into pYX142 (Novagen) via EcoRI/BamHI. The plasmid pYX142-AGC1-CTD was constructed by inserting the PCR product of yeast AGC1-CTD coding sequence (from nucleotide 1501–2709) into pYX142 via EcoRI. Transformations of W303 and Wagc1Δ strains were done by the improved lithium acetate method (Gietz and Schiestl, 1995). For the selection of geneticin (G418, Life Technologies) resistance, cells were spread on YPD plates containing 200 µg ml−1 G418 (Zúñiga et al., 1999) and for selection of plasmid transformants, MM plates (2% glucose) without leucine were used. Yeast sporulation and tetrad analysis of heterozygous deletants was performed using standard techniques and media. At least nine complete tetrads were dissected using an MSM Singer micromanipulator. Deletants were verified by PCR and Southern blot analysis.
Growth studies were initiated with late log precultures grown on minimal medium (MM) [0.17% yeast nitrogen base without amino acids and ammonium sulphate (YNB, Difco), 0.5% ammonium sulphate, 0.2% glucose as carbon source and 30 µg ml−1l-leucine, 20 µg ml−1l-tryptophan, 20 µg ml−1l-histidine, 20 µg ml−1 uracil and 20 µg ml−1 adenine]. The precultures were diluted with synthetic minimal medium (SM) (0.67% YNB, 0.12% ammonium sulphate, 0.1% KH2PO4, pH 4.5) containing different carbon sources and the appropriate supplements until a final optical density of 5 × 10−3 at 600 nm was reached. Growth studies in rich YP medium (1% yeast extract, 2% bactopeptone) with 2% glucose (YPD), 2% lactate (YPL), 0.5 mM oleic acid + 1% tergitol (YPO) and 100 mM Na-acetate (YPA) as carbon sources were started from medium log precultures grown on YPD and diluted with YPD, YPL, YPO or YPA to an optical density of 0.01 at 600 nm. When present, the concentration of aminooxyacetic acid was 3 mM. For isolation of mitochondria, yeast cells were precultured on MM and then diluted in SG [0.67% YNB, 0.1% casaminoacids (Difco), 2% galactose, and 20 mg l−1l-tryptophan, 40 mg l−1 adenine] at a final optical density of 0.3 at 600 nm (Arechaga et al., 1993).
Yagc1Δ cells were obtained by homologous recombination of the auxotrophic marker HIS3 at the AGC1 gene locus of S. cerevisiae YPH499 (Roussel et al., 2002). The YPH499 strain was transformed with the pYES2-AGC1 plasmid obtained by cloning the AGC1 coding sequence into the vector pYES2. Transformants were selected for uracil auxotrophy and precultured in synthetic complete medium [SC (Sherman, 1991)] lacking uracil and supplemented with 3% glycerol. For growth studies, the precultures were diluted with SC until a final optical density of 5 × 10−2 at 600 nm was reached. For preparation of mitochondria, the precultures were diluted 50-fold in YP medium containing 0.1% glucose as the Gal-Cyc promoter was repressed under these conditions. The pYES2-AGC1 cells were grown to exponential phase, and galactose (0.4%) was added 6 h before harvesting.
14 CO2 production from glutamate in yeast mitochondria
Yeasts were grown on SG until late log phase (A600 after 1:20 dilution, 0.4–0.7) (Bouillaud et al., 1994) and used to obtain mitochondrial preparations (Arechaga et al., 1993). Yeast mitochondria were suspended in 0.6 M mannitol, 2 mM EGTA, 10 mM Tris-maleate, 0.5 mM Na2HPO4, 0.2% bovine serum albumin, pH 6.8. To study glutamate or α-ketoglutarate metabolism, mitochondria (2 mg protein ml−1) were incubated in the same medium with either 0.1 mM l-[1-14C]glutamate (0.05 µCi ml−1) or 0.1 mM [1-14C]α-ketoglutarate (0.05 µCi ml−1). Mitochondrial suspensions (0.5 ml) were incubated in P24-plates (2.5 cm2) at room temperature, with mild orbital shaking, during 10–20 min, in the presence or absence of 5 mM aminooxyacetic acid. The evolved 14CO2 was trapped in filters (10 mm diameter; Millipore AP 250 1000) presoaked with 3.5 N NaOH and placed on top of the dishes when the experiment was started. At the end of the reaction time, 100 µl of 0.4 M citric acid was added to the mitochondrial suspension to stop the reaction and further trap the labelled carbon dioxide formed (45 min), and the radioactivity accumulated in the filters was counted. Appropriate correction was applied for radioactivity evolved in similar incubations performed without mitochondria.
The wild-type and Wagc1Δ strains were grown on SM with 2% sodium pyruvate as carbon source to late log phase (A600, 1.80–2.15) at 30°C, collected by centrifugation (8000 g for 2 min), washed twice in suspension medium (SM without supplements) and then suspended to a density of 1 g wet weight ml−1 in Na [2-13C] acetate or Na [1,2-13C] acetate 15–100 mM (Cerdán et al., 1990) or unlabelled acetate, where incubation was continued for different times. Incubations were stopped by freezing the yeast suspension in dry ice and subsequently at −70°C. Perchloric acid extracts were prepared by adding 10% HClO4 (Bogónez et al., 1983). After vigorous shaking, and three freeze/thaw cycles, the suspension was left on ice (1–2 h) and centrifuged (5000 g, 5 min). The supernatants were neutralized with ice-cold 5 M KOH, and after being kept on ice for 1–2 h, the precipitates were eliminated by centrifugation (5000 g, 5 min) and supernatants were lyophilized. These were then used for the acquisition of 13C NMR spectra of the extracts. Parallel samples were used for the determination of amino acids by automatic ion exchange chromatography with post-column ninhydrin derivatization.
Extracts for 13C NMR were re-suspended in 2H2O (99.9%2H, Apollo Scientific) and analysed under high resolution conditions by 13C NMR (90.55 MHz, 22°C, pH 7.2). Conditions were: π/3 pulses, 20 000 Hz spectral width, 32K computer memory corresponding to 0.6 s acquisition time, 6 s total cycle time and approximately 8000 scans. Broad band proton decoupling was applied only during the acquisition.
This work was supported by grants from the Spanish Ministerio de Ciencia y Tecnología (PM1998-0021), Ministerio de Sanidad (FISS 01/0395), Química Farmaceútica Bayer, S.A, the Ministero dell’Università e della Ricerca (MIUR-PRIN, MIUR-FIRB), the Centro di Eccellenza di Genomica comparata dell’Università di Bari (CEGBA), the European Social Fund, the Medical Research Council, UK and by an institutional grant from the Fundación Ramón Areces to the Centro de Biología Molecular Severo Ochoa.
The expert technical assistance of Fernando García Muñoz, from the Centro de Diagnóstico de Enfermedades Metabólicas, and that of Bárbara Sesé and Immaculada Ocaña is gratefully acknowledged. We are indebted to Dr E. Rial and Dr J. P. García Ballesta for their invaluable help and advice.
- 1993) Cysteine residues are not essential for uncoupling protein function. Biochem J 296: 693–700. , , , , , , et al. (
- 1998) Molecular cloning of Aralar, a new member of the mitochondrial carrier superfamily that binds calcium and is present in human muscle and brain. J Biol Chem 273: 23327–23334. , and (
- 2000) Characterization of a second member of the subfamily of calcium-binding mitochondrial carriers expressed in human non-excitable tissues. Biochem J 345: 725–732. , , and (
- 2002) Expression of the aspartate/glutamate mitochondrial carriers aralar1 and citrin during development and in adult rat tissues. Eur J Biochem 269: 3313–3320. , , , , , , and (
- 2001) Stoichiometry and compartmentation of NADH metabolism in Saccharomyces cerevisiae. FEMS Microbiol Rev 25: 15–37. , , , , , , and (
- 2002) Expression of three mitochondrial solute carriers, citrin, aralar1 and ornithine transporter, in relation to urea cycle in mice. Biochim Biophys Acta 1574: 283–292. , , , , , , et al. (
- 1983) Ammonia accumulation in acetate-growing yeast. Biochim Biophys Acta 733: 234–241. , , and (
- 1994) A sequence related to a DNA recognition element is essential for the inhibition by nucleotides of proton transport through the mitochondrial uncoupling protein. EMBO J 13: 1990–1997. , , , , , , et al. (
- 1970) Ca2+ metabolism in yeast cells and mitochondria. Biochim Biophys Acta 205: 18–26. , , , and (
- 1990) Cerebral metabolism of [1,2-13C2]acetate as detected by in vivo and in vitro 13C NMR. J Biol Chem 265: 12916–12926. , , and (
- 1996) Two yeast homologs of ECA39, a target for c-Myc regulation, code for cytosolic and mitochondrial branched-chain amino acid aminotransferases. J Biol Chem 271: 20242–20245. , , and (
- 1993) Abundant bacterial expression and reconstitution of an intrinsic membrane-transport protein from bovine mitochondria. Biochem J 294: 293–299. , , and (
- 1998) Expression in Escherichia coli, functional characterization, and tissue distribution of isoforms A and B of the phosphate carrier from bovine mitochondria. J Biol Chem 273: 22782–22787. , , and (
- 2001) Identification of the human mitochondrial oxodicarboxylate carrier. Bacterial expression, reconstitution, functional characterization, tissue distribution, and chromosomal location. J Biol Chem 276: 8225–8230. , , , , , , and (
- 2002) Identification of the mitochondrial glutamate transporter. Bacterial expression, reconstitution, functional characterization, and tissue distribution of two human isoforms. J Biol Chem 277: 19289–19294. , , , , , and (
- 1995) Transforming genes with DNA. Methods Mol Cell Biol 5: 255–269. , and (
- 1981) 13C NMR study of transamination during acetate utilization by Saccharomyces cerevisiae. Proc Natl Acad Sci USA 78: 2693–2697. , , and (
- 1970) The glutamate dehydrogenases of yeast: extra-mitochondrial enzymes. Biochim Biophys Acta 201: 13–19. , , and (
- 2003) Identification of a mitochondrial transporter for basic amino acids in Arabidopsis thaliana by functional reconstitution into liposomes and complementation in yeast. Plant J 33: 1027–1035. , , , , , and (
- 1970) Pathways of hydrogen in mitochondria of Saccharomyces carlsbergensis. Eur J Biochem 12: 583–592. , and (
- 1978) Arginine metabolism in Saccharomyces cerevisiae: subcellular localization of the enzymes. J Bacteriol 133: 1096–1107. , , and (
- 1998) Global regulatory functions of Oaf1p and Pip2p (Oaf2p), transcription factors that regulate genes encoding peroxisomal proteins in Saccharomyces cerevisiae. Mol Cell Biol 18: 6560–6570. , and (
- 1996) Mitochondrial and cytosolic branched-chain amino acid transaminases from yeast, homologs of the myc oncogene-regulated Eca39 protein. J Biol Chem 271: 24458–24464. , , , , and (
- 1999) The gene mutated in adult-onset type II citrullinaemia encodes a putative mitochondrial carrier protein. Nat Genet 22: 159–163. , , , , , , et al. (
- 1974) Evidence for electrogenic aspartate transport in rat liver mitochondria. Arch Biochem Biophys 161: 544–550. , , and (
- 1998) The importance of the glycerol 3-phosphate shuttle during aerobic growth of Saccharomyces cerevisiae. Yeast 14: 347–357. , , , , , and (
- 1998) The Saccharomyces cerevisiae NDE1 and NDE2 genes encode separate mitochondrial NADH dehydrogenases catalyzing the oxidation of cytosolic NADH. J Biol Chem 273: 24529–24534. , , , , , and (
- 2001) Central carbon metabolism of Saccharomyces cerevisiae explored by biosynthetic fractional (13) C labeling of common amino acids. Eur J Biochem 268: 2464–2479. , , , , , and (
- 1993) Evidence for three different electrophoretic pathways in yeast mitochondria: ion specificity and inhibitor sensitivity. J Bioenerg Biomembr 25: 671–678. , and (
- 2002) Identification and reconstitution of the yeast mitochondrial transporter for thiamine pyrophosphate. EMBO J 21: 5653–5661. , , , , , and (
- 1987) Isolation and expression of the gene encoding yeast mitochondrial malate dehydrogenase. J Bacteriol 169: 5157–5166. , and (
- 1992) AAT1, a gene encoding a mitochondrial aspartate aminotransferase in Saccharomyces cerevisiae. Biochim Biophys Acta 1171: 211–214. , , and (
- 2002) Kinetic regulation of the mitochondrial glycerol-3-phosphate dehydrogenase by the external NADH dehydrogenase in Saccharomyces cerevisiae. J Biol Chem 277: 27991–27995. , , , , , , and (
- 2003) The mitochondrial transporter family (SLC25): physiological and pathological implications. Eur J Physiol (in press). (
- 1995) Mitochondrial metabolite carrier proteins: purification, reconstitution, and transport studies. Methods Enzymol 260: 349–369. , , , and (
- 2000a) Yeast mitochondrial carriers: bacterial expression, biochemical identification and metabolic significance. J Bioenerg Biomembr 32: 67–77. , , , , and (
- 2000b) Identification and functions of new transporters in yeast mitochondria. Biochim Biophys Acta 1459: 363–369. , , , , , , et al. (
- 2001a) Citrin and aralar1 are Ca(2+)-stimulated aspartate/glutamate transporters in mitochondria. EMBO J 20: 5060–5069. , , , , , , et al. (
- 2001b) Identification in Saccharomyces cerevisiae of two isoforms of a novel mitochondrial transporter for 2-oxoadipate and 2-oxoglutarate. J Biol Chem 276: 1916–1922. , , , , , and (
- 2003) Developmental changes in the Ca(2+)-regulated mitochondrial aspartate-glutamate carrier aralar1 in brain and prominent expression in the spinal cord. Dev Brain Res 143: 33–46. , , , , , , et al. (
- 1995) The membrane of peroxisomes in Saccharomyces cerevisiae is impermeable to NAD(H) and acetyl-CoA under in vivo conditions. EMBO J 14: 3480–3486. , , , , and (
- 2002) Does any yeast mitochondrial carrier have a native uncoupling protein function? J Bioenerg Biomembr 34: 165–176. , , , , and (
- 1991) Getting started with yeast. Methods Enzymol 194: 3–21. (
- 1998) Identification of a cytosolically directed NADH dehydrogenase in mitochondria of Saccharomyces cerevisiae. J Bacteriol 180: 4051–4055. , and (
- 2002) Yeast mitochondrial oxodicarboxylate transporters are important for growth on oleic acid. Arch Biochem Biophys 406: 96–104. , , , , and (
- 1996) Mathematical models for determining metabolic fluxes through the citric acid and the glyoxylate cycles in Saccharomyces cerevisiae by 13C-NMR spectroscopy. Eur J Biochem 242: 770–778. , , , and (
- 1996) PCR-synthesis of marker cassettes with long flanking homology regions for gene disruptions in S. cerevisiae. Yeast 12: 259–265. (
- 1974) Interrelations between anion transport, ureogenesis and gluconeogenesis in isolated rat liver cells. In Regulation of Hepatic Metabolism. Lundquist, F., and Tygstrup, N. (eds). Copenhagen: Munksgaard, pp. 457–479. , , and (
- 1999) Disruption of six Saccharomyces cerevisiae novel genes and phenotypic analysis of the deletants. Yeast 15: 945–953. , , , , and (