Acetyl:Succinate CoA-transferase in Procyclic Trypanosoma brucei GENE IDENTIFICATION AND ROLE IN CARBOHYDRATE METABOLISM*

Acetyl:succinate CoA-transferase (ASCT) is an ace-tate-producing enzyme shared by hydrogenosomes, mitochondria of trypanosomatids, and anaerobically functioning mitochondria. The gene encoding ASCT in the protozoan parasite Trypanosoma brucei was identified as a new member of the CoA transferase family. Its as-signment to ASCT activity was confirmed by 1) a quantitative correlation of protein expression and activity upon RNA interference-mediated repression, 2) the absence of activity in homozygous (cid:1) asct / (cid:1) asct knock out cells, 3) (cid:3) Triton cell de- was removed by centrifugation for 5 min at 500 (cid:3) g , 4 The was determined by a radioactive assay as described previously (1). In brief, the assay contained 50 m M succinate, pH 7.4, 1 m M 14 C]acetyl-CoA (0.2 Mbq, Amersham Biosciences), 50 m M Tris-HCl, pH 7.4, m M MgCl , and 0.05% (v/v) Triton X-100. The assay was started by the addition of the homogenate (50–100 (cid:2) g protein), incubated at 20 °C for 10 min, and terminated by the addition of ice-cold trichloroacetic acid (10% final Reaction products were then separated by anion-exchange chromatography and quantified by liquid scintillation counting.

Acetyl:succinate CoA-transferase (ASCT) 1 is an enzyme that transfers the CoA moiety of acetyl-CoA to succinate, yielding acetate and succinyl-CoA. ASCT was earlier characterized as an acetate-producing activity in Trypanosomatidae (1). Trypanosomatidae are the earliest branching eukaryotes that retained mitochondria. They belong to the order Kinetoplastida, characterized by the kinetoplast, a highly concatenated form of the genome of the single mitochondrion. These unicellular eukaryotes contain unique organelles, such as glycosomes, which harbor part of the glycolytic pathway (2,3). The energy metabolism of Trypanosomatidae is characterized by excretion of mainly partially oxidized end products like pyruvate, succinate, and acetate. All of the Trypanosomatidae investigated produce acetate to a certain extent during their life cycle (4 -7), but the amount of acetate produced differs widely depending on the species and the stage of development.
Trypanosoma brucei, one of the causative agents of African trypanosomiasis, alternates during its life cycle between the bloodstream of its mammalian host and the blood-feeding tsetse insect vector Glossina spp. In the mammalian bloodstream long slender-form parasites proliferate. At the peak of parasitemia, nonproliferative short stumpy cells develop that are prepared to differentiate to procyclic insect stage parasites. The long slender stage depends entirely on glycolysis for energy generation and excretes pyruvate as the main end product of carbohydrate metabolism (8 -10). The procyclic stage, on the other hand, can also use amino acids for its energy generation, and its glucose metabolism is completely reorganized, with succinate and acetate as main end products. Recently it was shown that inside the glycosome succinate is produced by a soluble fumarate reductase (11). In the procyclic stage the mitochondrion is directly involved in the degradation of substrates, in contrast to the situation in the long slender bloodstream stage. Pyruvate enters the mitochondrion and is converted by pyruvate dehydrogenase into acetyl-CoA. This acetyl-CoA is not degraded to carbon dioxide via the Krebs cycle but is converted into acetate by ASCT (1,12). The ASCT reaction concomitantly produces succinyl-CoA, which is recycled to succinate by succinyl-CoA synthetase, an enzyme that also forms part of the Krebs cycle and produces ATP.
ASCT activity is also known to occur in the anaerobically functioning mitochondria of metazoa that produce acetate, such as parasitic helminths like Fasciola hepatica and Ascaris * The work in Bordeaux was supported by the CNRS, the Conseil Régional d'Aquitaine, the Groupement de Recherche CNRS-Parasitologie, the Ministère de l'Education Nationale de la Recherche et de la Technologie (Action Microbiologie), and the European Commission (International Cooperation-Developing Countries program). The work in Utrecht was supported by the Earth and Life Science Foundation (ALW) with financial aid from the Netherlands Organization of Scientific Research (NWO). The work in Munich was funded by the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ ¶ To whom correspondence may be addressed. suum (13,14). Furthermore, ASCT is also a key enzyme in the metabolism of a wide spectrum of anaerobic protists, including ciliates such as Nyctotherus ovalis, chytridiomycete fungi such as Neocallimastix, and parabasalids such as Trichomonas vaginalis (15)(16)(17). In these protists ASCT is located inside their hydrogenosomes, anaerobic energy-generating organelles. Hydrogenosomes are H 2 -producing, membrane-enclosed organelles related to mitochondria, and they evolved independently in the various protists (18,19). Sequence comparison of several organelle-specific heat shock proteins or chaperonins and the presence of targeting signals at the N terminus of hydrogenosomal enzymes that resemble mitochondrial import signals showed that hydrogenosomes and mitochondria are evolutionary related (20 -23). They probably originated from the same prokaryotic endosymbiont of the ␣ group of proteobacteria, and it is suggested that hydrogenosomes have evolved by adaptation to anaerobic conditions (24). The exact evolutionary relation between mitochondria and the various types of hydrogenosomes is, however, still debated.
Although the main function in energy generation of both mitochondria and hydrogenosomes is the oxidation of pyruvate and acetyl-CoA, the two types of organelles do not share many similarities. The ASCT/succinyl-CoA synthetase cycle is the only metabolic pathway common to mitochondria and hydrogenosomes (1). For that reason, phylogenetic analysis of ASCTs from different sources could provide a valuable tool to unravel the evolutionary history of the respective energy-generating organelles, the various types of hydrogenosomes, and (an)aerobically functioning mitochondria. Unfortunately, the sequence of not a single ASCT gene is known yet. Here we have identified the gene encoding ASCT of T. brucei. Several reverse genetic strategies and subcellular localization studies were used to establish the link between ASCT gene sequence and mitochondrial ASCT enzyme activity. Furthermore, two independent methods of metabolic analysis using 13 C-enriched and 14 C-labeled glucose were applied to wild type and transgenic mutant clones of T. brucei. The combined biochemical and genetic approach resulted in the gene identification of T. brucei ASCT and also demonstrated that this enzyme is indeed involved in acetate production. However, ASCT is not an essential gene as ASCT is not the only enzyme producing acetate in T. brucei.

EXPERIMENTAL PROCEDURES
Plasmid Constructs-The sheared DNA clone 7D10 from The Institute for Genomic Research (GenBank TM accessions AQ652285 and AQ652287) was obtained from Dr. Najib M. A. El-Sayed. The complete sequence of that part of chromosome 11 is now available from the TrypDB genome data base (Tb11.02.0290); it is identical with results of sequencing of clone 7D10. A MluI (blunted)/MscI fragment of the insert of 7D10 was excised and replaced in the reading frame orientation by a blasticidin S resistance BSD (25) cassette (SmaI/StuI fragment from pHD887 (courtesy of C. Clayton, Heidelberg) containing the PmlI fragment nucleotides 2189 -2687 from pcDNA6/V5-His C (Invitrogen) or by a phleomycin resistance BLE cassette (BamH1/NheI, both blunted, from pLew20 (26)) to generate the targeting vectors pBER2 and pBER3, respectively. Before transfection, the targeting vectors were linearized with BsrGI plus BsmI. To produce double-stranded RNA using head-to-head promoters, the Nterminal fragment of the T. brucei ASCT (nucleotides position 35 to position 665) was PCR-amplified from 7D10 and inserted into p2T7 (27) to give p2T7.ASCT. Plasmid pLew-ASCT-SAS contains a "sense/antisense" cassette targeting a 572-bp fragment of the T. brucei ASCT (nt position 85-657) in the pLew100 expression vector (kindly provided by E. Wirtz and G. Cross) (26) and was constructed as previously reported for repression of a flagellar protein (FTZC) gene expression (28). For overexpression in T. brucei, the complete ASCT-coding region was assembled from a N-terminal PCR product amplified from T. brucei 927/4 genomic DNA with the primers 5Ј-ggcagaattcCAGCTCCACATCACCTTCCAGG (EcoRI site) and 5ЈactgaagcttATGCTCCGCCGAACAAATTTT (HindIII site) and from a C-terminal fragment excised from plasmid 7D10 digested with EcoRI and BamHI. The two fragments were inserted into a HindIII/ BamHI-opened pTSArib (29). Southern blot analysis was done as described before (12) with an ASCT-specific probe (nucleotides 1-314 of the coding region) generated by PCR.
Production of Anti-ASCT Antibodies-A recombinant fragment corresponding to the first 414 amino acids of ASCT followed by a Cterminal histidine tag was expressed in the Escherichia coli BL21 using the pET16b expression vector (Novagen). After 2 h of induction at 37°C with 1 mM isopropyl-␤-D-thiogalactopyranoside, the cells were harvested by centrifugation, and proteins were purified by nickel chelate affinity chromatography according to the manufacturer's instructions (Novagen). Antisera were raised in rabbits by 3 injections at 15-day intervals of 200 g of recombinant purified protein, electro-eluted after separation on SDS-PAGE, and emulsified with complete (first injection) or incomplete Freund's adjuvant. The antiserum was affinity-purified on polyvinylidene difluoride membrane-bound recombinant protein (amino acids 12-493 fused to an N-terminal His tag in vector pQE32 (Qiagen)). Elution with 0.2 M glycine, 1 mM EGTA, pH 2.2, for 10 min was followed by immediate neutralization, stabilization in 200 g⅐ml Ϫ1 bovine serum albumin, and dialysis against PBS containing 0.02% NaNO 3 . One specific 53-kDa band was detected on the Western blot.
For immunofluorescence, log phase cells were fixed with formaldehyde as previously described (35). Slides were incubated with 1:100 diluted affinity-purified rabbit anti-ASCT, and the undiluted H95 monoclonal mouse anti-hsp60 (36) was followed by fluorescein isothiocyanate-conjugated goat anti-rabbit secondary antibody diluted 1:100 (Bio-Rad) and ALEXA Fluor 568-conjugated goat anti-mouse secondary antibody diluted 1:100 (Molecular Probes). Cells were observed with a Zeiss UV microscope, and images were captured by a MicroMax-1300Y/HS camera (Princeton Instruments) and MetaView software (Universal Imaging Corp.) and assembled in Adobe Photoshop.
Digitonin Fractionations-Digitonin fractionations were done as described in Saas et al. (37). Proteinase K digestion was adapted from Sveshnikova et al. (38). The pellet and concentrated supernatant were suspended in 500 l of trypanosome homogenization buffer without leupeptin (37), and 8 l of proteinase K (10 mg⅐ml Ϫ1 ) were added on ice. After 3 min at 37°C the reaction was stopped by adding 6 l of phenylmethylsulfonyl fluoride (100 mM diluted in isopropanol) on ice. For thermolysin digestion, the pellet and concentrated supernatant were suspended in 90 and 40 l, respectively, of 0.1 M Hepes-HCl, pH 7.2, 2 mM CaCl 2 , 0.2 g⅐ml Ϫ1 leupeptin, 0.6 M saccharose, 1 mM dithiothreitol. The pellet and supernatant were incubated with 10 and 5 l, respectively, of thermolysin (1 mg⅐ml Ϫ1 ) for 30 min at 45°C. The digestion was stopped by adding EDTA on ice.
ASCT Activity Assay-Homogenates of procyclic T. brucei were prepared from 5 ϫ 10 7 cells in 20 mM HEPES buffer, pH 7.4, containing 1% Triton X-100 using a Teflon glass homogenizer. Subsequently, cell debris was removed by centrifugation for 5 min at 500 ϫ g, 4°C. The activity of ASCT was determined by a radioactive assay as described previously (1). In brief, the assay mix contained 50 mM succinate, pH 7.4, 1 mM [1-14 C]acetyl-CoA (0.2 Mbq, Amersham Biosciences), 50 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , and 0.05% (v/v) Triton X-100. The assay was started by the addition of the homogenate (50 -100 g protein), incubated at 20°C for 10 min, and terminated by the addition of ice-cold trichloroacetic acid (10% w/v, final concentration). Reaction products were then separated by anion-exchange chromatography and quantified by liquid scintillation counting.
Immunoprecipitation-Procyclic cells (1 ϫ 10 8 cells) were harvested by centrifugation at 500 ϫ g for 5 min. The cells were washed twice and resuspended in 20 mM Tris-HCl buffer, pH 7.8, containing 1% Triton X-100. Antiserum was added to the cell lysate (20% v/v) and incubated on ice for 2 h. Subsequently, protein A-conjugated acrylic beads (Sigma, 150 M) in 20 mM Tris-HCl buffer, pH 7.8, were added to the lysate and rotated for 2 h at 4°C. The antigen-bead preparation was pelleted by centrifugation (500 ϫ g for 10 s). The supernatant was collected, and the pelleted beads were washed once with 20 mM Tris-HCl buffer, pH 7.8, and then resuspended in the same buffer. Both the supernatant and the resuspended beads were assayed for ASCT activity. Cell lysates treated with protein-A acrylic beads only were used as the control.
Nuclear Magnetic Resonance (NMR) Experiments-4 ϫ 10 9 T. brucei procyclic cells grown in the SDM79 medium (up to 1 ϫ 10 7 cells⅐ml Ϫ1 ) were incubated in 10 ml of incubation buffer (PBS buffer supplemented with 24 mM NaHCO 3 , pH 7.3) containing 110 mole D-[1-13 C]glucose (11 mM) for 90 -180 min at 27°C as described before (11,39). The cell viability and motility, checked every 30 min, were not affected during the incubation time. D-Glucose concentration in the medium was determined at the beginning and the end of the incubation with the D-glucose Trinder kit (Sigma). The supernatant was lyophilized and re-dissolved in 500 l of D 2 O, and 15 l of pure dioxane was added as an external reference. 13 C NMR spectra were collected as described before (11,39). The relaxation delay was 8 s for a nearly complete longitudinal relaxation. The specific 13 C enrichment of lactate (carbon C3), acetate (carbon C2), and succinate (carbons C2 and C3) was determined from 1 H-observed/ 13 C-edited NMR ( 1 H/ 13 C NMR) spectra acquired under 13 C decoupling (40, 41) as described before (39). The reproducibility and accuracy of the method were assessed using pure solutions of succinate, acetate, lactate, malate, pyruvate, or ␤-hydroxybutyrate; the relative errors on the 13 C enrichment determinations were Ͻ5%. 2 The amount of 13 C-enriched molecules (including D-glucose) was calculated on the basis of the dioxane peak. For each experiment, the amount of remaining D-glucose calculated from the NMR spectra corresponded Ϯ5-25% to the concentration measured with the D-glucose Trinder kit.
Metabolic Incubations-Proliferating T. brucei cells (starting at 5 ϫ 10 6 ml Ϫ1 ) were incubated for 18 -24 h at 27°C in sealed 25-ml Erlenmeyer flasks containing 5 ml of SDM79 incubation medium. Before sealing, the flasks were flushed for 1 min with a gas phase of air and CO 2 (95/5%). The incubations were performed after the addition of 5 Ci of D-[6-14 C]glucose. Incubations were terminated by the addition of 40 l of 6 M HCL to lower the pH to 3.5. Preceding acidification, 0.1 ml 1 M NaHCO 3 was added through the rubber stopper, and the flasks were placed on ice. Immediately after acidification, the incubation flasks were flushed with nitrogen for 90 min at 0°C. In this way all carbon dioxide was removed, whereas acetate remained in the incubation medium. The carbon dioxide was trapped in a series of four scintillation vials, each filled with 1 ml of 0.3 M NaOH and 15 ml of Tritisol scintillation fluid modified according to the method described by Pande (42,43). After removal of carbon dioxide, the acidified supernatant was separated from the cells by centrifugation (4°C, 10 min at 500 ϫ g) and neutralized by the addition of 40 l of 6 M NaOH. Analysis of the labeled end products occurred by anion-exchange chromatography on a Dowex 1X8, 100 -200 mesh column (Serva), 60 ϫ 1.1 cm in chloride form (44). The column was eluted successively with 200 ml of 5 mM HCl, 130 ml of 0.2 M NaCl, and 130 ml of 0.5 M NaCl. The fractions were collected and counted with 2 ml of Lumac LCS in a scintillation counter. All values were corrected for blank incubations. Labeled end products were identified by their R F values.
Glucose concentrations were determined enzymatically using a standard procedure (45). Protein was determined with a Lowry method using defatted and dialyzed bovine serum albumin (Roche Applied Science) as the standard (46).

Identification of an ASCT Candidate
Gene-The transfer of the coenzyme A (CoA) group is catalyzed by a large family of prokaryotic and eukaryotic CoA transferases (47) that share conserved protein domains (CoA_trans in protein families data base, accession PF01144) and have a broad range of substrate specificities. Acetoacetate, a preferred substrate of mammalian succinyl-CoA:3-ketoacid-CoA transferase, is structurally comparable with acetate, the substrate of ASCT. Therefore, we first searched for succinyl-CoA:3-ketoacid-CoA transferase-related gene sequences in the available T. brucei genome databases using BLAST. Only a single high scoring segment pair was identified, and the target shares up to 55% amino acid identity over its full length with succinyl-CoA:3-2 M. Biran and P. Canioni, unpublished data.  (37) as internal reference (two color detection with the Odyssey system). B, double plot representing relative protein levels (normalized mean Ϯ S.D. of three independent Western blots) and relative ASCT enzyme activities. Uninduced activities were between 61 and 80 nmol⅐min Ϫ1 ⅐mg Ϫ1 protein.
Protein levels and activities were expressed as % of the matched uninduced sample. The limit of sensitivity of the Western method was determined in parallel by dilution series to be Յ3%. C, relative amounts (% of total) of major products of glucose breakdown. Tetracycline-induced (filled symbols) and uninduced (open symbols) cells were labeled with [6-14 C]glucose for 24 h before harvest and analysis by ion exchange chromatography. Labeled CO 2 and other minor products were less than 5% of the total end products. ketoacid-CoA transferase from several vertebrate species. A single copy (see Fig. 3 and data not shown) of this gene maps to chromosome 11 of the T. brucei genome reference strain TREU 927/4 (see TrypDB, release 2.1, systematic name Tb1102.0290). The gene encodes a 53-kDa protein (493 amino acids) with two CoA transferase domains in tandem, the hallmark of eukaryotic CoA transferases (48). All nine amino acids found in immediate proximity of the catalytic center in the crystal structure of porcine succinyl-CoA:3-ketoacid-CoA transferase (48) are conserved in the new CoA transferase identified in T. brucei. Orthologues were also identified in T. cruzi (76% amino acid identity, GeneDB_Tcruzi) and Leishmania major (65% amino acid identity, GeneDB_Lmajor).
RNAi-mediated Repression of ASCT Activity-To correlate expression of the ASCT candidate gene with ASCT activity, we generated procyclic T. brucei cell lines for inducible RNAi-mediated repression of the protein. A fragment of the candidate open reading frame was placed between inverted copies of a tetracycline-inducible T7 promoter and was stably introduced into the host line 29-13 clone 6 (transgenic for the tetracycline repressor and T7 polymerase) for production of double-stranded RNA (2p3-1 cell line). At different time points after induction with tetracycline, the ASCT activity was measured as described before (1) in lysates of uninduced and induced 2p3-1 cells. The specific activity declined up to 10-fold over 96 h (Fig. 1B). Rabbit antibodies against the candidate open reading frame were produced after recombinant expression in E. coli (see "Experimental Procedures") and were used for accurate quantitation of the protein by Western blotting followed by detection with infrared fluorescent probes (Odyssey system, Licor). The time course of RNAimediated repression shows an excellent correlation between the amount of the 53-kDa protein and ASCT activity (Fig. 1). After 96 h the level of the protein was at or below the limit of detection (3% of uninduced, determined from dilution series). The slow decline over four days indicated a relatively long half-life. Very similar data (Fig. 2) were also obtained with two independent clones (hp3 and hp4), resulting from a different strategy to produce double-stranded RNA, namely expression of an inverted repeat of the target sequence that folds into an RNA hairpin (construct pLew-ASCT-SAS). Immunoreactive protein and ASCT activity were both reduced 10-fold after 96 h of tetracycline induction. Note that the procyclic host lines and RNAi vectors are different in the two experiments shown in Figs. 1 and 2. Together, these independent data sets strongly suggest that the candidate gene encoded ASCT. An important function of ASCT was anticipated from the observed growth phenotype. The population doubling times of RNAi clones hp3 and hp4 increased from 10.7 to 15.4 h and from 11.1 to 15.3 h, respectively, upon tetracycline induction.
Gene Deletion and Overexpression of T. brucei ASCT-To confirm the identity of the ASCT gene, procyclic lines with targeted gene deletion of ASCT were generated. Replacement vectors with phleomycin and blasticidin resistance markers (BLE and BSD, respectively) were constructed and transfected, and correct homologous integration of the targeting vectors in the resulting drug resistant clones was analyzed by PCR using two independent primer pairs per locus type (wt, ⌬asct::BLE, ⌬asct::BSD). Whereas targeting of the first allele resulted in correct integration in 5 out of 5 clones, only 3 out of 52 independent double resistant clones were confirmed as homozygous gene deletions and were verified by Southern analysis (Fig. 3,  A and B). This unusually low frequency of homologous integration and the low transfection efficiency in those experiments indicated strong selection against loss of the last ASCT gene copy. This interpretation was also supported by a severe growth phenotype of the three independent ⌬asct::BLE/⌬asct::BSD knock out clones (Fig. 4). The initial population doubling time increased from 11 h (wt) to 18 h (clone 45), 27 h (clone 3.7), and 30 h (clone 3.23). Also, a higher fraction of aberrant cell divisions was revealed by microscopic scoring of 4Ј6-diamidino-2phenylindol-stained nucleus/kinetoplast configurations in those clones (not shown). The absence of ASCT protein on Western blots (Fig. 3C) and a 4 -5-fold reduced ASCT activity in knock out clones (Fig. 3D) confirmed deletion of ASCT. The significant residual enzyme activity measured by the ASCT assay in ⌬asct::BLE/⌬asct::BSD knock out clones was due to an acetate-producing activity that is independent of succinate and, hence, distinct from ASCT (data not shown).
For genetic rescue of the ⌬asct/⌬asct phenotype and stable overexpression of ASCT, we introduced into clone 45 (⌬asct::BLE/⌬asct::BSD) the construct pTSArib.ASCT, which integrates into the ribosomal promoter region. Quantitative Western analysis of dilution series documented 16-fold overexpression (Fig. 3C). This resulted in a 3-fold overexpression of ASCT activity (Fig. 3D), again confirming the correlation between ASCT expression and specific enzyme activity. Partially inactive protein or posttranslational regulation of the activity may explain the disproportionality between ASCT protein level and enzyme activity upon overexpression. The rescued line had a normal population distribution of nucleus/kinetoplast configurations (not shown), but only partial reversion of the growth phenotype was achieved (Fig. 4). Because 8-fold overexpression in the wild type background resulted in a normal growth rate (not shown), the incomplete reversion of the growth phenotype of clone 45 is likely to be due to the selection history rather than to overexpression of ASCT.
Immunoprecipitation using the antiserum against the recombinant protein fragment (see "Experimental Procedures") resulted in depletion of ASCT activity from cell lysates of wild type procyclic cells (98 Ϯ 3% depletion) and of the transgenic clone overexpressing ASCT (79 Ϯ 14% depletion). No activity could be detected in the resuspended acrylic beads fractions, which is probably due to binding of the polyclonal antibodies to the protein.
Mitochondrial Localization of ASCT-ASCT activity had been previously localized to mitochondria of procyclic T. brucei (1). Therefore, we determined the subcellular localization of the cloned ASCT gene product with two different methods, immunofluorescence microscopy and digitonin fractionation. Fig. 5 shows that anti-ASCT antibodies specifically stain a tubular structure, readily identified as the mitochondrion by the hsp60 marker in wt cells but not in ⌬asct/⌬asct cells or in induced RNAi knockdown cells. Wild type procyclic cells were also fractionated by increasing concentrations of digitonin, and ASCT protein was quantified by Western blotting in the soluble and insoluble fractions (Fig. 6). Established cytosolic (phosphoglycerate kinase) and mitochondrial (hsp60) markers were run as internal controls. The fractionation behavior of ASCT was virtually identical to that of hsp60, indicating a mitochondrial matrix localization. This was confirmed by digestion of selected digitonin fractions with two different proteases, proteinase K (not shown) and thermolysin (Fig. 6) before Western analysis. ASCT remained fully protected in the mitochondrial pellet fraction but was completely digested in the soluble fraction. In agreement with this localization, a short (11-12 amino acids) N-terminal mitochondrial target sequence was predicted by algorithms available via internet services (MitoProt II, TargetP, iPSORT) and by sequence alignment to vertebrate succinyl-CoA:3-ketoacid-CoA transferase precursors. The colocalization of ASCT activity (1) and the candidate gene product in the mitochondrion strengthens our evidence for identification and cloning of the bona fide ASCT gene.
Analysis of Glucose Metabolism after Manipulation of the T. brucei ASCT Expression-We used a carbon 13 ( 13 C) NMR spectroscopy analysis to detect and quantify the metabolic end products excreted by the wild type and mutant cell lines. The parasites were incubated in PBS/NaHCO 3 medium containing 11 mM D-[1-13 C]-glucose as the only carbon source until up to half of D-glucose was consumed by each cell line. The incubation medium was then analyzed by NMR spectroscopy. The wild type AnTat1.1 and EATRO1125 procyclic cells mainly excrete 13 C-enriched succinate, acetate, and lactate (ϳ70, 18, and 7% of the 13 C-enriched excreted molecules, respectively) with traces of malate, fumarate, and alanine ( Fig. 7A and Table  I) (11,39). Surprisingly, the rate of acetate excretion in the AnTat1.1 and derived ⌬asct::BLE/⌬asct::BSD (KO) cell lines was in the same range (122 and 132-136 nmol of 13 C-enriched excreted molecules⅐h Ϫ1 ⅐mg Ϫ1 protein, respectively), whereas the rate of succinate excretion increased by 17-66% in the KO cell lines (Table I). Consequently, the total amount of 13 Cenriched end products excreted by the KO cell lines increased by 21-65%. To interpret these data the rate of glucose con- sumption was considered. The KO cell lines consumed significantly more glucose compared with the parental AnTat1.1 (Table I). This was measured during the incubation for NMR analysis (30 -40% increase) as well as in separate incubations in SDM79 medium (47-55% increase). The ratio of acetate/ succinate excretion was significantly reduced in the KO cell lines as compared with AnTat1.1 cells (10.9 -14.7 versus 16.4% of the 13 C-enriched excreted molecules, respectively). Thus, deletion of the ASCT gene resulted in a relative reduction of acetate production, although other activities must account for maintenance of acetate excretion in the absence of ASCT (see "Discussion"). The KO cell lines but not the wild type also excreted detectable amounts of 13 C-enriched pyruvate and ␤-hydroxybutyrate (Fig. 7, A-B).
The rescued ⌬asct/⌬asct cell line showed rates of glucose consumption and end product excretion similar to AnTat1.1, and no excretion of pyruvate and ␤-hydroxybutyrate was detected (Fig. 7C). As expected for overexpression of ASCT protein and activity (Fig. 3), the amount of excreted acetate significantly increased from 122 in the wild type to 188 nmol of FIG. 4. Long term growth of ASCTdeficient and rescued cell clones. Cells were cultured in SDM79 medium over 47 days and were regularly passaged by dilution to maintain the density between 5 ϫ 10 5 and 1 ϫ 10 7 cells⅐ml Ϫ1 . The cumulative cell number is plotted in log scale on the y axis. The genotype of clones 45, 3.7, and 3.23 is ⌬asct::BLE/⌬asct::BSD; the genotype of the clone 45-derived rescued line is ⌬asct::BLE/⌬asct::BSD ASCT HYG. The culture time corresponding to the number of population doublings at the start of radioactive glucose incubations and NMR analysis is indicated.  Excreted end products (numbers in Fig. 7) Percentage of 13 13 C-enriched excreted molecules⅐h Ϫ1 ⅐mg Ϫ1 protein in the rescued line (Table I). Clearly, the metabolic phenotype of ⌬asct/ ⌬asct cells was fully rescued by extragenic ASCT expression, confirming the genetic link between the metabolic phenotype and ASCT activity. Very similar metabolic changes were measured in tetracycline-induced RNAi hp3 and hp4 cells when compared with the parental EATRO1125-T7T cell line, i.e. an increase of the rates of glucose consumption and total end product excretion, a slight decrease of the ratio of acetate and succinate production, and excretion of some ␤-hydroxybutyrate (Table I). The non-induced RNAi hp3 and hp4 control cell displayed an intermediate metabolic phenotype. This was not surprising, as a certain leakiness of the tetracycline-inducible expression systems in T. brucei, which drives transcription of double-stranded RNA, is generally observed (26,28,39). Indeed, we quantified the amount of ASCT protein in all non-induced RNAi clones and found 50% of the protein level measured in the matched wild type cell lines (data not shown).
In addition to the 13 C NMR studies, which could only be performed on harvested cells at high cell density in a buffered salt medium, we also investigated the end product formation in growing wild type and mutant procyclic cells under more physiological conditions. The major radioactive end products excreted during growth in SDM79 culture medium containing 6-14 C-labeled glucose were acetate and succinate as reported before (12). The total radioactivity of minor unidentified me-tabolites that were detected in these experiments never exceeded 10% of the sum of acetate and succinate. Therefore, only acetate and succinate are represented in Table II. The absence of ASCT did not result in a major change in the ratio of acetate and succinate produced, although a slight but not significant decrease in acetate production is suggested by the comparison of the wild type and ⌬asct/⌬asct KO 45 cells (Table II). This is in agreement with the [ 14 C]glucose incubations performed with RNAi clone 2p3.1 (Fig. 1) and also with the 13 C NMR analysis of end products excreted by ⌬asct/⌬asct clones and RNAi clones hp3 and hp4 (Table I). The estimated [ 14 C]glucose degradation (calculated from the sum of excreted products) was not increased in KO 45 cells compared with the wild type parent (Table II). It should be noted that the experimental conditions of [ 14 C]glucose incubations and 13 C NMR analysis are very different and that KO 45 showed the weakest metabolic phenotype of several ⌬asct/⌬asct clones (see Table I).
Analysis of the KO 45 rescue line that overexpresses ASCT activity (see Fig. 3) showed a significant increase in acetate production, whereas succinate production was decreased (Table II). This is again in agreement with the 13 C NMR studies (Table I) and provides direct evidence for an important role of ASCT in acetate production under physiological conditions. DISCUSSION ASCT enzyme activity in T. brucei is encoded by a newly identified member of the eukaryotic CoA transferase gene family. This is the first report of a gene encoding ASCT. Several lines of evidence support the gene-to-activity assignment. 1) The completed genome sequence of T. brucei contains only one member of the highly conserved and well recognizable CoA transferase gene family I (47). 2) A quantitative correlation was documented between ASCT activity and the amount of protein encoded by the candidate ASCT gene. Inducible RNAi-mediated repression with two independent methods was used.
3) Homozygous-targeted deletion of the ASCT gene resulted in protein product and activity levels below the experimental background. 4) Overexpression of the candidate gene product in T. brucei resulted in increased ASCT activity and increased acetate production. 5) Both the candidate gene product and ASCT activity are expressed in procyclic form T. brucei but not in the bloodstream stage of the parasite (data not shown). 6) Mitochondrial localization of the candidate ASCT gene product was verified by two independent methods and corresponds to mitochondrial localization (1) of the ASCT activity. 7) ASCT FIG. 5. Subcellular localization of ASCT by immunofluorescence analysis. Control procyclic EATRO1125-T7T cells, the tetracycline (tet)-induced clone RNAi hp3 (RNAi ϩtet) and ⌬asct/⌬asct clone 45 (KO) were stained with rabbit anti-ASCT (fluorescein channel (top panels)) and mouse anti-hsp60 (Alexa 568 channel (middle panels)). Phase contrast images of the same cells are shown in the bottom panels.
FIG. 6. Localization of ASCT by subcellular fractionation. Procyclic AnTat1.1 were fractionated by increasing digitonin concentrations. A, soluble (S) and mitochondrial pellet (P) fractions were subjected to Western blotting and probed with affinity-purified rabbit anti-ASCT antibodies, mouse monoclonal anti-hsp60, and rabbit anti-phosphoglycerate kinase (PGK) C serum (cross-reacting with phosphoglycerate kinase B, courtesy of P. Michels) followed by simultaneous two-color infrared fluorescence detection in the Odyssey system. B, quantitative representation of the data from A. The % release of a specific protein into the supernatant was calculated from the ratio of band intensities S/SϩP and plotted against the log of digitonin concentration. C, digestion of selected soluble (S) and mitochondrial pellet (P) fractions with thermolysin protease before Western analysis. activity could be removed by immunoprecipitation with a specific antiserum raised against the candidate ASCT gene product. Unfortunately, heterologous expression of ASCT protein in E. coli, Saccharomyces cerevisiae, and Pichia pastoris resulted in insoluble protein with no detectable activity.
In procyclic T. brucei, acetyl-CoA produced from glucose is converted into acetate and is not metabolized through the Krebs cycle (12). Van Hellemond et al. (1) previously characterized a mitochondrial ASCT activity in trypanosomatids involved in acetate production from acetyl-CoA. Here, we have investigated the role of ASCT for acetate production in wild type and genetically engineered trypanosomes by qualitative and quantitative analyses of the excreted end products of glucose metabolism. To our surprise the level of acetate excretion was only slightly reduced in the mutant cell lines devoid of ASCT activity, suggesting an additional pathway(s) involved in acetate production from glucose. In fact, a minor succinateindependent acetate production from acetyl-CoA was observed before (1). This activity could be increased in the ⌬asct/⌬asct cell lines as an adaptive change taking place during selection of the knock out lines. Although acetate excretion is only slightly reduced, the mutant cell lines show a distinct metabolic phenotype of ASCT deficiency. 1) The ⌬asct/⌬asct cell lines excrete pyruvate and ␤-hydroxybutyrate, which are not detectable in the incubation medium of the wild type. In mammals, ␤-hydroxybutyrate is the end product of a four step pathway used to eliminate excess of acetyl-CoA. Because possible orthologues of the respective enzymes are present in the T. brucei genome sequence (data not shown), a ␤-hydroxybutyrate-producing pathway is likely to be present in this parasite. In addition, pyruvate is the precursor of the ASCT substrate acetyl-CoA. Thus, the excretion of both ␤-hydroxybutyrate and pyruvate may be the consequence of acetyl-CoA accumulation in those mutant cells, which are less efficient in converting acetyl-CoA into acetate due to the absence of ASCT. Our interpretation is that the alternative acetate producing pathway(s) may not suffice to metabolize all the produced acetyl-CoA. 2) The rate of glucose consumption significantly increased as a consequence of ASCT gene deletion or inactivation of ASCT gene expression (14 -63% increase, depending on the cell line and the experimental conditions). Succinyl-CoA produced by ASCT can be converted back into succinate by the succinyl-CoA synthetase with a net production of ATP by substrate level phosphorylation (1). The increase of glucose consumption in the absence of ASCT, thus, may be an adaptive compensation of reduced energy yield from glucose. 3) The RNAi and KO mutant cell lines show significant reduction of their growth rate, which demonstrates that ASCT activity is important for normal functioning of procyclic T. brucei. The metabolic phenotype was fully rescued, and the growth phenotype was partially rescued by extragenic ASCT expression.
Together, the phenotype of ASCT deficiency, which seems to be due to accumulation of acetyl-CoA, and the fact that overexpression of ASCT resulted in a major increase of acetate excretion in proliferating parasites, leave no doubt that ASCT is a major source for acetate production in the normal situation. Although ASCT is not an essential gene and an alternative acetate-producing pathway(s) exists, the importance of the ASCT for physiological functioning is clearly shown. 1) The recovery of drug resistant clones upon targeting of the second allele was very low, and 2) we observed that in most of these drug-resistant clones the targeting vector was not correctly integrated in the ASCT locus, a very rare event in T. brucei, where homologous recombination is the rule. Although ⌬asct/ ⌬asct cells are viable, short term adaptation does not overcome the phenotype discussed above. Upon continuous passage for an extended period, ⌬asct/⌬asct trypanosomes are finally able to recover the wild type growth rate (Fig. 4). The basis of this long term adaptation is not known.
In conclusion, our genetic approaches, the localization studies, and the two independent metabolic studies all show that the identified gene codes for the ASCT enzyme in T. brucei. Acetate production via ASCT occurs in many parasites (protozoa as well as helminths) but is absent in their vertebrate hosts. The enzyme seems, therefore, to be a logical target for broad-spectrum anti-parasitic drugs. This strategy will not be useful to combat African trypanosomiasis, because ASCT is not detected in the bloodstream stage of T. brucei infecting the mammalian host. ASCT is, however, present in infective stages of other trypanosomatids including Leishmania. The gene se-quence of T. brucei ASCT will facilitate identification of ASCT in anaerobic protists such as N. ovalis, Neocallimastix, and T. vaginalis, which contain the various types of hydrogenosomes. Phylogenetic studies on ASCT, one of the few enzymes shared by hydrogenosomes and mitochondria of some species, could shed more light on the evolutionary relationships between mitochondria and the various types of hydrogenosomes.