Acetate Produced in the Mitochondrion Is the Essential Precursor for Lipid Biosynthesis in Procyclic Trypanosomes

Acetate produced in the mitochondrion is the essential precursor for lipid biosynthesis in procyclic trypanosomes

Loı¨cRivie`rea, Patrick Moreaub, Stefan Allmannc, Matthias Hahnc, Marc Birand, Nicolas Plazollesa, Jean-Michel Franconid, Michael Boshartc, and Fre´de´ ric Bringauda,d,1

aLaboratoire de Microbiologie Cellulaire et Molećulaire et Pathogeńicite´, Unite´Mixte de Recherche 5234 Centre National de la Recherche Scientifique, bLaboratoire de Biogene`se Membranaire, Unite´Mixte de Recherche 5200 Centre National de la Recherche Scientifique, and dRe´sonance Magne´tique des Syste`mes Biologiques, Unite´Mixte de Recherche 5536, Centre National de la Recherche Scientifique, Universite´Victor Segalen Bordeaux 2, 146 rue Leó Saignat, 33076 Bordeaux cedex, France; and cBiozentrum, Department Biologie I, Genetik, Ludwig-Maximilians-Universita¨t Mu¨ nchen, D-82152 Planegg-Martinsried, Germany

Edited by M. Daniel Lane, Johns Hopkins University School of Medicine, Baltimore, MD, and approved June 11, 2009 (received for review March 27, 2009) Acetyl-CoA produced in mitochondria from carbohydrate or amino synthesis of fatty acids (10) are noteworthy. Indeed, whereas acid catabolism needs to reach the cytosol to initiate de novo most cells use either a type I or type II synthase for de novo fatty synthesis of fatty acids. All eukaryotes analyzed so far use a acid synthesis, T. brucei is the only known organism using a third citrate/malate shuttle to transfer acetyl group equivalents from the mechanism for this process. This parasite uses microsomal mitochondrial matrix to the cytosol. Here we investigate how this elongases to synthesize short-, medium-, and long-chain fatty acetyl group transfer occurs in the procyclic life cycle stage of acids from butyryl-CoA, by using malonyl-CoA as a C2 donor, Trypanosoma brucei, a protozoan parasite responsible of human whereas other organisms only can extend premade medium and sleeping sickness and economically important livestock diseases. long FAs by elongases to make long and very long fatty acyl Deletion of the potential citrate lyase gene, a critical cytosolic chains (11). In addition, we show here that the procyclic try- enzyme of the citrate/malate shuttle, has no effect on de novo panosomes also use a pathway to produce cytosolic acetyl-CoA to feed de novo biosynthesis of fatty acids. biosynthesis of fatty acids from 14C-labeled glucose, indicating that In nonphotosynthetic eukaryotes, nearly all of the acetyl-CoA another route is used for acetyl group transfer. Because acetate is used in fatty acid synthesis is formed in mitochondria from produced from acetyl-CoA in the mitochondrion of this parasite, pyruvate oxidation and from the catabolism of the carbon we considered genes encoding cytosolic enzymes producing skeletons of amino acids. It is widely accepted that all eukaryotes acetyl-CoA from acetate. We identiﬁed an acetyl-CoA synthetase analyzed so far use the citrate/malate shuttle to transfer acetyl gene encoding a cytosolic enzyme (AceCS), which is essential for group equivalents from the mitochondrial matrix to the cytosol cell viability. Repression of AceCS by inducible RNAi results in a across the acetyl-CoA-impermeable mitochondrial inner mem- 20-fold reduction of 14C-incorporation from radiolabeled glucose brane [see (12)]. In this shuttle, intramitochondrial acetyl-CoA or acetate into de novo synthesized fatty acids. Thus, we demon- first reacts with oxaloacetate to form citrate, in the citric acid strate that the essential cytosolic enzyme AceCS of T. brucei is cycle (TCA) reaction catalyzed by citrate synthase (CS). Citrate responsible for activation of acetate into acetyl-CoA to feed de passes through the mitochondrial inner membrane on a citrate novo biosynthesis of lipids. To date, Trypanosoma is the only transporter or a citrate/malate exchanger, and is converted back known eukaryotic organism that uses acetate instead of citrate to to acetyl-CoA and oxaloacetate by the cytosolic citrate lyase transfer acetyl groups over the mitochondrial membrane for cyto- (CL) (see Fig. 1). In trypanosomes, existence of a citrate/malate solic lipid synthesis. shuttle was assumed to meet the needs of cytosolic fatty acid biosynthesis, although this view is not supported by any exper- acetyl-CoA synthetase ͉ citrate/malate shuttle ͉ de novo lipid imental data (11, 13). Here we demonstrate that this assumption biosynthesis ͉ mitochondrial acetate is incorrect for the procyclic stage of T. brucei. Instead, this parasite expresses a cytosolic ‘‘AMP forming’’ acetyl-CoA synthetase (AceCS), which is essential for incorporation of radio- Trypanosoma brucei is an unicellular eukaryote, belonging to the labeled glucose (or acetate) into fatty acids and essential for protozoan order Kinetoplastida, that causes sleeping sickness in growth. Thus, acetyl-CoA is converted into acetate in the humans and economically important livestock diseases. This mitochondrion (7), which cross the mitochondrial inner mem- parasite undergoes a complex life cycle during transmission from brane to be converted into acetyl-CoA by the cytosolic AceCS. the bloodstream of a mammalian host (bloodstream stages of the We have demonstrated that acetate can be used to exchange parasite) to the alimentary tract (procyclic stage) and salivary acetyl group equivalents between the mitochondrial and cyto- glands (epimastigote and metacyclic stages) of a bloodfeeding solic compartments. insect vector, the tse-tse fly. In addition to their relevance for health and development in subsaharan Africa, trypanosomes are Results famous for a variety of very unusual genetic and biochemical A Citrate/Malate Shuttle Is Not Required for Lipid Biosynthesis. It was features that stimulate broad scientific and evolutionary interest. assumed by analogy (8) that acetyl-CoA is exchanged in try- These include exotic mechanisms of gene expression like poly- cistronic transcription of genes (1), maturation of premessenger Author contributions: L.R., P.M., M. Boshart, and F.B. designed research; L.R., P.M., S.A., RNA by trans-splicing and extensive editing of mitochondrial M.H., M. Biran, N.P., and F.B. performed research; J.-M.F. contributed new reagents/analytic RNAs (2, 3), and sophisticated mechanisms of immune evasion tools; L.R., P.M., S.A., M.H., M. Biran, M. Boshart, and F.B. analyzed data; and L.R., P.M., M. like antigenic variation and antibody endocytosis (4, 5). In the Boshart, and F.B. wrote the paper. context of this report, metabolic peculiarities like the compart- The authors declare no conﬂict of interest. mentalization of glycolysis in glycosomes, which are peroxisome- This article is a PNAS Direct Submission. like organelles (6), the presence of a single, developmentally 1To whom correspondence should be addressed. E-mail: [email protected]. regulated mitochondrion per cell with unusual enzyme activities This article contains supporting information online at www.pnas.org/cgi/content/full/ (7, 8), energy metabolism (9) and unusual pathways for de novo 0903355106/DCSupplemental.

12694–12699 ͉ PNAS ͉ August 4, 2009 ͉ vol. 106 ͉ no. 31 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903355106 Downloaded by guest on September 28, 2021 Glucose Glycosomes Mitochondrion

Succinate Succinate -1 cells)

Mal TCA KG 8 G3P DHAP 2 Cit

1 4 CPM • (10 PEP Pyr AcCoA Acetate 5 6 1 23 4 1 23 4 1 23 4 1 23 4 1 23 4 7 PC PI PE TAG FFA Fatty acids AcCoA Acetate Fig. 2. D-[U-14C]-glucose incorporation into lipids of the EATRO1125.T7T 3 Cit and ⌬cl/⌬cl cell lines. This ﬁgure shows the relative amounts of 14C-labeled phospholipids (PC, PI, and PE) and 14C-labeled neutral lipids (TAG and FFA) Lipids extracted from the EATRO1125.T7T (black columns labeled 1) and ⌬cl/⌬cl knock-out (gray columns labeled 2, 3, and 4 for ⌬cl/⌬cl-C9, ⌬cl/⌬cl-E9, and Fig. 1. Schematic representation of lipid biosynthesis from D-glucose me- ⌬cl/⌬cl-B6, respectively) cell lines incubated in the presence of D-[U-14C]- tabolism in the procyclic form of T. brucei. Black arrows represent enzymatic glucose. Error bars indicate m Ϯ SD of 3 independent experiments. steps of D-glucose metabolism, and excreted end products are boxed. Dashed arrows indicate steps, which are supposed to occur at a background level or not at all. The enzymes targeted by RNAi are indicated by a boxed number. De which is a major source of acetyl-CoA in standard growth novo lipid biosynthesis is schematically represented by double line arrows. The glycosomal compartments, the mitochondrial compartment and the tricar- conditions (15)). Under these conditions, acetyl-CoA and glyc- 14 boxylic acid cycle (TCA) are indicated. Abbreviations: AcCoA, acetyl-CoA; Cit, erol-3-phosphate (G3P) produced from D-[U- C]-glucose ca- citrate; DHAP, dihydroxyacetone phosphate; G3P, glycerol 3-phosphate; KG, tabolism (Fig. 1) are major precursors for label incorporation 2-ketoglutarate; MAL, malate; PEP, phosphoenolpyruvate; Pyr, pyruvate. In- into the fatty acid and glycerol parts, respectively, of the glyc- dicated steps are: 1, citrate synthase (CS); 2, citrate transporter or citrate/ erolipids. Indeed, L-proline is a poor acetyl-CoA producer and malate exchanger; 3, citrate lyase (CL); 4, acetate:succinate CoA-transferase does not produce G3P in the presence of D-glucose (16). The (ASCT); 5, unknown enzyme; 6, passive diffusion or acetate carrier; 7, ‘‘AMP- major fatty-acid containing lipids labeled de novo were the forming’’ acetyl-CoA synthetase (AceCS). phospholipids (PC, phosphatidylcholine; PE, phosphatidyleth- anolamine; PI, phosphatidylinositol) and the neutral lipids (TAG, triacylglycerols; FFA, free fatty acids). No significant panosomes by a putative citrate/malate shuttle between the 14 mitochondrion, where it is produced, and the cytosol, where it change in C-incorporation into individual lipids was seen in the ⌬ ⌬ feeds de novo lipid synthesis. This shuttle requires 2 citrate 3 cl/ cl clones compared with parental cells (Fig. 2), which forming/consuming enzymes (the mitochondrial CS and the clearly demonstrated that the conventional citrate/malate shuttle cytosolic CL) and a citrate/malate exchanger or a citrate trans- is not essential for fatty acid biosynthesis in procyclic trypano- porter present in the mitochondrial inner membrane (Fig. 1). somes. Whereas putative CS and CL genes have been identified in trypanosomatid genomes (8), a functional role of a citrate shuttle A Functional Acetyl-CoA Synthetase (AceCS) Gene in T. brucei. With remains to be proven. To investigate the role of the putative CL the aim of characterizing new acetate consuming or producing gene (Tb927.8.6820, http://www.genedb.org/genedb/tryp/) in enzymes in T. brucei, we searched for sequences related to lipid biosynthesis of procyclic trypanosomes, we have generated acetaldehyde dehydrogenase (EC: 1.2.1.3), acetyl-CoA hydro- 3 independent CL knock-out mutant cell lines (⌬cl/⌬cl-B6, -C9, lase (EC: 3.1.2.1), phosphate acetyltransferase (EC: 2.3.1.8), BIOCHEMISTRY and -E9) by replacing both CL alleles by selectable markers in acetate kinase (EC: 2.7.2.1), ‘‘ADP-forming’’ acetyl-CoA syn- the procyclic AnTat1.1–1313 cell line. Deletion of both alleles thetase (EC: 6.2.1.13), and ‘‘AMP-forming’’ acetyl-CoA syn- was confirmed by PCR (Fig. S1B) and RT-PCR (Fig. S1C). thetase (EC: 6.2.1.1) in the GeneDB T. brucei genome database These mutant clones are viable, although their population (http://www.genedb.org/genedb/tryp/) using TBLASTN. Only doubling time is increased (16.6, 16.7, and 14.6 h, respectively) the Tb927.8.2520 sequence encoding a protein homologous to compared with the AnTat1.1–1313 parental cell line (12.5 h). ‘‘AMP-forming’’ acetyl-CoA synthetase (AceCS) showed a sig- The genetic data are compatible with negative results obtained nificant match to the query sequences. The 674 amino acid with SB-204990, the cell-penetrant ␥-lactone prodrug of the protein encoded by the T. brucei Tb927.8.2520 gene shows potent ATP citrate-lyase inhibitor SB-201076 (Ki ϭ l ␮M for 49.6%, 49.2%, and 47.7% identity with the Drosophila melano- human enzyme) [(14), kindly provided by SmithKline Beecham gaster, Arabidopsis thaliana, and human AceCS. Pharmaceuticals]. Up to 100 ␮M did not induce any effect on The T. brucei AceCS gene was expressed in E. coli with an growth of procyclic trypanosomes of wild type and ⌬asct/⌬asct N-terminal polyhistidine tag, using the pET vector system (No- knock-out cell lines [reduced acetate production (7), see below vagene). The recombinant AceCS protein (75 kDa) was purified ϩ and Table S1], whereas 30 ␮M almost completely inhibited lipid to near homogeneity on Ni2 nitriloacetic acid-agarose. A synthesis in cultured mammalian cells (12, 14). In T. brucei, specific activity of 0.3 units (mmol⅐minϪ1) per mg of protein was toxicity of SB204990 was only observed after 68 h of incubation determined in the direction of acetyl-CoA formation. For each with Ͼ200 ␮M. It should be noted that we have no proof that of the 3 substrates involved in the acetyl-CoA production, the SB204990 uptake, prodrug conversion and CL inhibition are Michaelis-Menten curves were hyperbolic with apparent Km effective in T. brucei. To directly investigate a role of the putative values of 80 ␮M for acetate, 400 ␮M for ATP, and 20 ␮M for CL gene in lipid biosynthesis, a quantitative analysis of 14C- CoASH (Fig. S2). These values are similar to those obtained for labeled lipids produced from D-[U-14C]-glucose was performed other AceCS, including the human enzyme (17–19). on the ⌬cl/⌬cl and parental cell lines. The parasites were incubated in SDM80 medium containing D-[U-14C]-glucose and AceCS Is an Essential Cytosolic Protein in Procyclic Trypanosomes. A L-proline as the main carbon sources (without L-threonine, rabbit immune serum raised against the recombinant AceCS

Riviere et al. PNAS ͉ August 4, 2009 ͉ vol. 106 ͉ no. 31 ͉ 12695 Downloaded by guest on September 28, 2021 Fig. 3. Subcellular localization of AceCS by digitonin titration. The presence of AceCS, FHc, PPDK, and ASCT was determined by western blot analysis in the supernatant (S) and pellet (P) fractions from cells incubated with 0.02–0.6 mg digitonin/mg protein in STE buffer containing 150 mM NaCl, as indicated. The digitonin concentrations required to fully release cytosolic (Cyto), glycosomal (Glyco), and mitochondrial (Mito) marker proteins are indicated by an arrow- head below the blots.

protein recognizes a single 75-kDa protein expressed in the procyclic form of T. brucei (see Fig. 3). The subcellular localization was determined by a digitonin titration experiment, wherein the different membranes of the procyclic trypanosomes were differentially permeabilized by increasing concentrations of the detergent. A western blot analysis of the soluble and insoluble fractions indicates that AceCS and the cytosolic fu- marase (FHc) (20) are released together at 40 ␮g digitonin/mg protein, whereas the glycosomal PPDK (21) and the mitochondrial ASCT (7) are released at much higher digitonin concentrations (200 ␮g and 400 ␮g of digitonin per mg protein, respectively) (Fig. 3). Immunofluorescence analysis of the T. brucei procyclic form using the anti-AceCS immune serum confirmed a homogenous diffuse pattern characteristic for a cytoplasmic localization. We used the pLew100 vector (22) for inducible RNAi- mediated degradation of the AceCS transcript (23), as described (24, 25). A recombinant pLew-AceCS-SAS plasmid designed to produce, under the control of tetracycline, hairpin double- stranded RNA molecules specific for AceCS was introduced into the EATRO1125.T7T cell line expressing the tetracycline re- pressor and T7 RNA polymerase (22, 24). We generated 2 mutant cell lines (RNAiAceCS-1 and RNAiAceCS-2), that effi- ciently repress AceCS expression upon tetracycline induction of RNAi (Fig. 4A, Lower). AceCS is essential for the procyclic cells, RNAi Fig. 4. Inactivation of AceCS gene expression by RNAi. A shows growth because both induced AceCS mutant cell lines die 18 days curves of the parental cell line maintained in the presence (EATRO1125.T7T ϩ after tetracycline addition (Fig. 4A). Nocodazole) or the absence (EATRO1125.T7T) of nocodazole (1.2 ␮g/mL), and 2 independent RNAi clones, RNAiAceCS-1 and RNAiAceCS-2. Cells were main- AceCS Is Not Involved in Acetate Production. Because AceCS can tained in the exponential growth phase (between 106 and 107 cells per mL) by theoretically function in both directions (acetate or acetyl-CoA dilution, and cumulative cell numbers are displayed. ‘‘Induced’’ or ‘‘non production), we first investigated its role in acetate production. induced’’ refers to tetracycline addition. Below the growth curve a western The rate of 13C-enriched acetate produced from D-[1-13C]- blot of an induction time course from 1 to 15 days is shown for the RNAiAceCS-1 RNAi glucose metabolism by the RNAiAceCS-1 cell line is not signifi- cell line. The hsp60 protein serves as internal control. In B ( AceCS-1, tetracycline induction) and C (EATRO1125.T7T, nocodazole addition) the cantly changed upon tetracycline induction (Table S1). It has 14C-labeled phospholipids (PC, PI, PE) and the neutral lipids (NL) were sepa- been previously shown that acetyl-CoA produced from D- rated by HPTLC and analyzed as described in the Methods section. [1-14C]- glucose metabolism in procyclic trypanosomes does not enter the acetate was added 16 h before lipid extraction. TCA cycle but is rather converted into acetate that is excreted (between 20% and 40% of the excreted end products) (25, 26). In procyclic trypanosomes, ASCT is involved in acetate produc- conditions, we generated the RNAiASCT/AceCS double mutant tion from D-glucose metabolism (27). However, the rate of cell line for tetracycline-inducible repression of both ASCT and acetate production by ⌬asct/⌬asct and induced RNAiASCT cell AceCS proteins (Fig. S3). Carbon 13 (13C) NMR spectroscopy lines is only moderately reduced (approximately one-third com- was used to compare the metabolic end products excreted by 2 pared with parental cell lines, see Table S1), suggesting that independent single RNAiASCT-cl3 and RNAiASCT-cl4 mutant another unknown enzymatic activity is involved in acetate clones (7) and the double RNAiASCT/AceCS mutant cell line. The production (7). AceCS was an obvious candidate. To uncover a relative amount of 13C-enriched acetate produced from D-[1- possible effect on acetate production under more challenging 13C]-glucose metabolism is equivalent (14%, 14.6%, and 14.2%

12696 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903355106 Riviere et al. Downloaded by guest on September 28, 2021 13 of the C-enriched excreted end products, respectively) in all 3 AB lines after 8–12 days of RNAi-induction (Table S1). These data suggest that AceCS is not significantly contributing to acetate -1 production and strongly support the view that AceCS functions -1

in the direction of acetyl-CoA production from acetate. cells) cells) 8 8

AceCS Provides the Acetyl-CoA Precursor for Lipid Biosynthesis. The T. brucei AceCS thus is an essential cytosolic enzyme that seems CPM • CPM (10 to produce acetyl-CoA from acetate. To investigate its role in de • CPM (10 RNAi novo fatty acid biosynthesis, the tetracycline-induced - 1 23 4 1 23 4 1 23 4 1 23 4 1 23 4 1 23 4 AceCS-1 mutant and the parental cell line were incubated in PC PI PE TAG FFA Esters SDM80 medium (devoid of L-threonine) containing [1-14C]- Fig. 5. D-[U-14C]-glucose incorporation into lipids of the EATRO1125.T7T acetate. De novo biosynthesis of phospholipids PC, PI, PE, and and RNAiAceCS-1 cell lines. A shows 14C-labeled phospholipids (PC, PI, and PE) neutral lipids is shown in Fig. 4 B and C. Label incorporation into and 14C-labeled neutral lipids (TAG and FFA), respectively, separated by HPTLC lipids rapidly decreased after tetracycline addition, and lipid and analyzed as described in the Methods section. 14C-labeled fatty acid esters biosynthesis was almost completely abolished after 7 days of released from saponiﬁcation of total lipids are presented in B. Black columns induction (Fig. 4B). Cells were still alive at this time point and correspond to the EATRO1125.T7T parental cell line, whereas gray columns cell death occurred approximately 18 days after induction (Fig. correspond to the noninduced (column 2), tetracycline-induced for 5 days 4A). The same results were obtained with the RNAiAceCS-2 cell (column 3), or tetracycline-induced for 12 days (column 4) RNAiAceCS-1 cells. RNAi D-[U-14C]-glucose was added 16 h before lipid extraction. Error bars indicate line. It is noteworthy that the energy metabolism of the - Ϯ AceCS-1 cell line is not significantly affected, because the m SD of 3 independent experiments. amount of excreted end products from D-[1-13C]-glucose metabolism is similar before or 9–10 days after tetracycline- human cell lines showing 40-fold difference in acetyl-CoA incubation (935 versus 912 nmol/h/mg of protein, Table S1). To synthetase (AceCS) activity (17), indicating that AceCS is not exclude that the observed decrease of lipid biosynthesis was not involved in lipid biosynthesis from glucose. an indirect consequence of growth phenotype (Fig. 4A), we used Here, we show that an eukaryotic cell, the procyclic stage of a control with similar growth properties instead of unin- Trypanosoma brucei, uses acetate instead of citrate to transfer ducedRNAiAceCS-1 cells. Parental cells were grown in the pres- acetyl group equivalents across the mitochondrial membranes. ence of the cytoskeleton inhibitor nocodazole, which affects cell In this transfer system, acetate is produced in the trypanosome division. Whereas 1.2 ␮g/mL nocodazole produced a growth mitochondrion from acetyl-CoA by ASCT (7, 27) and an alter- curve similar to tetracycline induction of the RNAiAceCS cell lines native unknown step that is currently under investigation. The (Fig. 4A), de novo lipid biosynthesis from [1-14C]-acetate was alternative acetate production step is probably mitochondrial, essentially unaffected (Fig. 4C). This strongly supported our because acetate production is abolished in the tetracycline- hypothesis that AceCS is providing most of the acetyl-CoA for RNAi de novo lipid biosynthesis. induced PDH-E2 cell line lacking the mitochondrial pyru- Because acetate is not a physiological carbon source of vate dehydrogenase complex (16). Then acetate crosses the trypanosomes, the RNAiAceCS-1 cell line was also incubated in mitochondrial inner membrane by passive diffusion [acetate SDM80 medium containing L-proline and D-[U-14C]-glucose. being a lipophilic weak acid in its protonated form (28)] or Both lipid precursors, glycerol-3-phosphate and acetyl-CoA, are through an unknown mitochondrial carrier or exchanger. Once produced from D-[U-14C]-glucose. Therefore we compared 14C- in the cytosol, acetate is converted into acetyl-CoA by the label incorporation into glycerolipids (PC, PI, PE, and TAG ‘‘AMP-forming’’ AceCS. The essential role of AceCS in de novo derived from both precursors) with free fatty acids, which only lipid biosynthesis was demonstrated in this report by the 20-fold contain the 14C-label from acetyl-CoA (Fig. 1). The amounts of decrease of radiolabel incorporation into fatty acids from the BIOCHEMISTRY free fatty acids produced by the tetracycline-induced RNAi- glucose precursor when AceCS expression was inhibited by AceCS-1 cell line after 12 days of induction are 20-fold reduced RNAi. Very little is known about lipid requirements of procyclic compared with the parental cell line (Fig. 5A). This cell line also trypanosomes, however, supplementation with lipids or FBS is RNAi shows a 16-fold reduction of fatty acid esters obtained by total essential for optimal growth (29). The AceCS cell lines die lipid saponification (Fig. 5B), confirming that glucose-derived approximately 18 days after RNAi induction, suggesting that de acetyl-CoA is provided by AceCS. Although glycerolipids (PC, novo lipid biosynthesis is essential and not complemented by PI, PE, and TAG) are also labeled from the G3P precursor, it is uptake of lipids present in the serum-containing medium. The noteworthy that the labeling of these lipids was significantly citrate/malate shuttle is apparently not functional, at least in the reduced in the induced RNAiAceCS-1 cell line (4.8, 2.5, 4.2, and procyclic stage of the parasite, because no rescue of the RNAi- 1.8-fold after 12 days of tetracycline induction), consistent with AceCS phenotype is observed. strong reduction of 14C-labeled fatty acid incorporation in these In other model organisms, AceCS is an ubiquitous cytosolic classes of lipids (Fig. 5A). and/or mitochondrial enzyme, whose main role is to activate the lipophilic acetate into the metabolic precursor acetyl-CoA (30). Discussion The single AceCS gene in the T. brucei genome encodes a The first enzyme in the de novo synthesis pathway of fatty acids, cytosolic enzyme catalyzing exactly this biochemical reaction. the acetyl-CoA carboxylase, is located in the cytosol. Therefore, However, the substrate of the cytosolic T. brucei AceCS (acetate) acetyl-CoA produced in mitochondria from carbohydrate or is produced in the mitochondrion, whereas in other organisms amino acid catabolism needs to exit this organelle. All eu- acetate is produced in the same subcellular compartment or is karyotes analyzed so far use the citrate/malate shuttle, involving taken up from the extracellular environment. For example, in E. CS and CL, to export acetyl-CoA from mitochondria for supply coli and other prokaryotes, glucose is first converted into acetate of cytosolic lipid biosynthesis. This is exemplified by reduction of (and eventually excreted). When glucose becomes limiting, D-[6-14C]-glucose-dependent lipid biosynthesis in human cells by extracellular acetate is assimilated and activated into acetyl-CoA RNAi inhibition of CL expression or chemical inhibition of CL by AceCS to feed gluconeogenesis and the TCA cycle (31). In the activity (SB-204990) (12). In contrast, no significant difference cytosol of Saccharomyces cerevisiae and other yeasts, pyruvate was observed in 14C-pyruvate-dependent lipid biosynthesis in 2 produced from glycolysis can be converted into acetaldehyde

Riviere et al. PNAS ͉ August 4, 2009 ͉ vol. 106 ͉ no. 31 ͉ 12697 Downloaded by guest on September 28, 2021 and acetate, which is activated into acetyl-CoA by the cytosolic Methods AceCS; this pathway is a bypass of the conventional pyruvate Trypanosome Culture and Transfection. The procyclic form of the dehydrogenase complex also expressed in yeasts (32). The yeast EATRO1125.T7T and AnTat1.1–1313 strain was cultivated at 27°C in SDM79 or cytosolic AceCS, also translocated to the nucleus, is the source SDM80 media supplemented with 10% (vol/vol) of heat-inactivated FCS, and ␮ ␮ ␮ of acetyl-CoA for the nuclear histone acetyltransferases (33). In 7.5 g/mL hemin, in the presence of 10 g/mL neomycine and 25 g/mL hygromycin (EATRO1125.T7T) or 15 ␮g/mL neomycine and 2.5 ␮g/mL phleo- mammals, there are several occasions in which acetate must be mycin (AnTat1.1–1313) (15). Procyclic AnTat1.1 1313 were grown in SDM79 as activated by AceCS to be metabolized, however, as opposed to above plus 10 mM glycerol, 23.8 mM NaHCO3 and 100 U/mL penicillin, and trypanosomes, acetate is never produced in mitochondria. For selected in 2.5 ␮g/mL phleomycin, 10 ␮g/mL blasticidin, and 1 ␮g/mL puromy- instance, (i) the liver activates acetate produced by bacteria in cin. The parental EATRO1125 and AnTat1.1 cell lines were derived from the the colon (34) or (ii) produced from ethanol degradation in the same trypanosome stock (40). Transfection of procyclic trypanosomes was cytosol, (iii) the cytosolic acetyl-CoA hydrolase activities pro- performed as described before (7). duce significant amount of acetate, which can be salvaged by Inhibition of Gene Expression by RNA Interference and Gene Knock-Out. The AceCS to reenter the metabolism (35), and (iv) in neurons, the inhibition by RNAi of the expression of the ASCT and AceCS genes in the cytosolic acetylcholinesterases breakdown the neurotransmitter EATRO1125.T7T procyclic form of T. brucei was performed as previously acetylcholine into choline and acetate that must be reactivated described (7, 22–25). See SI Methods for more details. Replacement of the CL by AceCS to replenish acetylcholine stores (36). gene by the blasticidin and puromycin resistance markers via homologous The T. brucei genome contains putative or characterized genes recombination was performed by transfecting the AnTat1.1–1313 procyclic for all of the key enzymes of the citrate/malate shuttle, including cell line with PCR amplified resistance markers using long primers homologous over 50 bp to CL UTR sequences. Correct homologous integration of the CS and CL, the mitochondrial and cytosolic malate dehydroge- resistance markers in the resulting drug resistant clones was analyzed by PCR nases and a putative tricarboxylate exchanger (37), however the (see Fig. S1). RNA was isolated from 109 cells by using 5 mL of TriZol LS putative CL genes is not essential as shown by knock-out strains (Invitrogen). RNA integrity was checked on an agarose gel and quantified (Fig. 2). In contrast, the tetracycline-induced RNAiAceCS cell using a nanodrop. Total RNA (5 ␮g) was DNase I-treated by using the TURBO lines show an almost complete absence of 14C-incorporation into DNA-freeTM kit (Ambion) and converted to cDNA by using oligo(dT)12–18 lipids when D-[U-14C]-glucose is the main carbon source and die primers (Invitrogen) and SuperScript™ II Reverse Transcriptase. Primers used for RT-PCR are shown in Table S2. after 18 days, indicating that CL and consequently the citrate/ malate shuttle are not used in procyclic trypanosomes to feed Production of Recombinant AceCS and Specific Polyclonal Antibodies. For lipid biosynthesis. In addition, we have no proof that the expression of the recombinant T. brucei AceCS in E. coli and AceCS activity test, Tb927.8.6820 gene product is a bona fide CL gene. First, the CL see SI Methods. activity is not detectable in extracts from procyclic and bloodstream forms of T. brucei; second, the Tb927.8.6820 gene (an- Lipid Labeling from [1-14C]-acetate and [U-14C]-glucose. Aliquots of 108 cells in notated as CL gene in GeneDB) shows a low level of similarity the late exponential phase were incubated for 16 h in 5 mL of modified SDM80 with CL from other organisms (the best hit in GenBank was 27% medium without L-threonine (15) containing 10 mM D-glucose, 1 mM acetate, and 10 ␮Ci of [1-14C]-acetate (58.9 mCi/mmol) or containing 3 mM D-glucose identity with the Roseobacter litoralis CL); and third, the MITOP and 5 ␮Ci of D-[U-14C]-glucose (287 mCi/mmol). Cells were checked microscop- program detected a signal peptide with high probability in the ically for viability several times during the incubation. Subsequently, lipids products of the Tb927.8.6820 gene and the orthologous T. cruzi were extracted by chloroform:methanol (2:1, vol/vol) and analyzed as de- and Leishmania braziliensis genes (0.7, 0.96, and 0.99, respec- scribed before (41, 42). See SI Methods for more details. tively), suggesting that these proteins are expressed in the mitochondrion. This observation and the moderate growth Digitonin Permeabilization. Digitonin titration was performed as previously phenotype of ⌬cl/⌬cl lines not explained by any change of described (43). See SI Methods for more details. 14C-incorporation into lipids from D-[U-14C]-glucose suggest an Western Blot Analyses. Total protein lysates of wild type or mutant procyclic alternative function of the Tb927.8.6820 gene that is currently trypanosomes (5 ϫ 106 cells) were separated by SDS/PAGE (12%) and blotted under investigation. on Immobilon-P™ (Millipore). Immunodetections were performed using as Although the citrate/malate shuttle is apparently not func- primary antibodies mouse anti-AceCS (1:100), rabbit anti-ASCT (1:500) (7), tional, we speculate that mitochondrial citrate production is mouse anti-FHc (1:100) (20), and rabbit anti-PPDK (1:100) (21). As secondary important for other reasons. A significant flux through the CS antibodies, anti-mouse or anti-rabbit IgG conjugated to horseraddish perox- reaction is indicated by 100-fold citrate accumulation in aconi- idase (Bio-Rad) were used. The ECL kit was used according to the manufac- turer’s instructions (GE Healthcare). tase knock-out mutant lines (26). The presence of a cytosolic aconitase activity (38) and the recent identification of a glyco- Nuclear Magnetic Resonance Experiments. 13C-enriched end products excreted somal NADP-dependent isocitrate dehydrogenase (39) suggest by parental and RNAi mutant cell lines from D-[1-13C]-glucose metabolism was that citrate produced in the mitochondrion might be transferred performed as described before (20, 25). See SI Methods for more details. to the cytosol, where it can be converted to isocitrate by the cytosolic aconitase activity. Once transferred to the glycosomes, ACKNOWLEDGMENTS. We thank SmithKline Beecham Pharmaceuticals Ltd. by a putative isocitrate/2-ketoglurate carrier (Tb10.389.0690) for providing us with the SB204990 prodrug and M. Berenguer for her contribution to CL inhibition experiments during undergraduate research train- present in the glycosomal proteome (39), isocitrate could feed ing in the Munich laboratory. This work was supported by the National Council the glycosomal NADP-dependent isocitrate dehydrogenase to for Scientific Research, the Universite´Victor Segalen Bordeaux 2, the Agence produce NADPH required by all oxidative stress defense sys- Nationale de la Recherche program METABOTRYP, the Fondation pour la Recherche Me´dicale in Bordeaux, the University of Munich, Deutsche For- tems. The relevance of this alternative glycosomal NADPH- schungsgemeinschaft SSP1131 in Munich, and the Franco-Bavarian University producing pathway is under investigation. Cooperation Center Grant BFHZ/CCUFBU (to M. Boshart and F.B.).

1. Ivens AC, et al. (2005) The genome of the kinetoplastid parasite, Leishmania major. 5. Engstler M, et al. (2007) Hydrodynamic ﬂow-mediated protein sorting on the cell Science 309:436–442. surface of trypanosomes. Cell 131:505–515. 2. Liang XH, Haritan A, Uliel S, Michaeli S (2003) trans and cis splicing in trypanosomatids: 6. Michels PA, Bringaud F, Herman M, Hannaert V (2006) Metabolic functions mechanism, factors, and regulation. Eukaryot Cell 2:830–840. of glycosomes in trypanosomatids. Biochim Biophys Acta 1763:1463– 3. Stuart KD, Schnaufer A, Ernst NL, Panigrahi AK (2005) Complex management: RNA 1477. editing in trypanosomes. Trends Biochem Sci 30:97–105. 7. Rivie`re L, et al. (2004) Acetyl:succinate CoA-transferase in procyclic Trypanosoma 4. Pays E (2005) Regulation of antigen gene expression in Trypanosoma brucei. Trends brucei. Gene identiﬁcation and role in carbohydrate metabolism. J Biol Chem Parasitol 21:517–520. 279:45337–45346.

12698 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903355106 Riviere et al. Downloaded by guest on September 28, 2021 8. van Weelden SW, van Hellemond JJ, Opperdoes FR, Tielens AG (2005) New functions 26. Van Weelden SW, et al. (2003) Procyclic Trypanosoma brucei do not use Krebs cycle for parts of the Krebs cycle in procyclic Trypanosoma brucei, a cycle not operating as activity for energy generation. J Biol Chem 278:12854–12863. a cycle. J Biol Chem 280:12451–12460. 27. Van Hellemond JJ, Opperdoes FR, Tielens AG (1998) Trypanosomatidae produce ace- 9. Bringaud F, Rivie`re L, Coustou V (2006) Energy metabolism of trypanosomatids: tate via a mitochondrial acetate:succinate CoA transferase. Proc Natl Acad Sci USA Adaptation to available carbon sources. Mol Biochem Parasitol 149:1–9. 95:3036–3041. 10. Lee SH, Stephens JL, Paul KS, Englund PT (2006) Fatty acid synthesis by elongases in 28. Wolfe AJ (2005) The acetate switch. Microbiol Mol Biol Rev 69:12–50. trypanosomes. Cell 126:691–699. 29. Coppens I, Courtoy PJ (1995) Exogenous and endogenous sources of sterols in the 11. Lee SH, Stephens JL, Englund PT (2007) A fatty-acid synthesis mechanism specialized for culture-adapted procyclic trypomastigotes of Trypanosoma brucei. Mol Biochem Para- parasitism. Nat Rev Microbiol 5:287–297. sitol 73:179–188. 12. Hatzivassiliou G, et al. (2005) ATP citrate lyase inhibition can suppress tumor cell 30. Starai VJ, Escalante-Semerena JC (2004) Acetyl-coenzyme A synthetase (AMP forming). growth. Cancer Cell 8:311–321. Cell Mol Life Sci 61:2020–2030. 13. van Hellemond JJ, Bakker BM, Tielens AG (2005) Energy metabolism and its compart- 31. Kumari S, et al. (2000) Regulation of acetyl coenzyme A synthetase in Escherichia coli. mentation in Trypanosoma brucei. Adv Microb Physiol 50:199–226. J Bacteriol 182:4173–4179. 14. Pearce NJ, et al. (1998) The role of ATP citrate-lyase in the metabolic regulation of 32. Boubekeur S, et al. (1999) A mitochondrial pyruvate dehydrogenase bypass in the yeast plasma lipids. Hypolipidaemic effects of SB-204990, a lactone prodrug of the potent ATP citrate-lyase inhibitor SB-201076. Biochem J 334:112–119. Saccharomyces cerevisiae. J Biol Chem 274:21044–21048. 15. Lamour N, et al. (2005) Proline metabolism in procyclic Trypanosoma brucei is down- 33. Takahashi H, McCaffery JM, Irizarry RA, Boeke JD (2006) Nucleocytosolic acetyl- regulated in the presence of glucose. J Biol Chem 280:11902–11910. coenzyme a synthetase is required for histone acetylation and global transcription. Mol 16. Coustou V, et al. (2008) Glucosed-induced remodelling of intermediary and energy Cell 23:207–217. metabolism in procyclic Trypanosoma brucei. J Biol Chem 283:16342–16354. 34. Scheppach W, Pomare EW, Elia M, Cummings JH (1991) The contribution of the large 17. Luong A, Hannah VC, Brown MS, Goldstein JL (2000) Molecular characterization of intestine to blood acetate in man. Clin Sci (Lond) 80:177–182. human acetyl-CoA synthetase, an enzyme regulated by sterol regulatory element- 35. Crabtree B, Souter MJ, Anderson SE (1989) Evidence that the production of acetate in binding proteins. J Biol Chem 275:26458–26466. rat hepatocytes is a predominantly cytoplasmic process. Biochem J 257:673–678. 18. Fujino T, Kondo J, Ishikawa M, Morikawa K, Yamamoto TT (2001) Acetyl-CoA syn- 36. Carroll PT (1997) Evidence to suggest that extracellular acetate is accumulated by rat thetase 2, a mitochondrial matrix enzyme involved in the oxidation of acetate. J Biol hippocampal cholinergic nerve terminals for acetylcholine formation and release. Chem 276:11420–11426. Brain Res 753:47–55. 19. Brasen C, Schonheit P (2005) AMP-forming acetyl-CoA synthetase from the extremely 37. Berriman M, et al. (2005) The genome of the African trypanosome Trypanosoma halophilic archaeon Haloarcula marismortui: Purification, identification, and expres- brucei. Science 309:416–422. sion of the encoding gene, and phylogenetic affiliation. Extremophiles 9:355–365. 38. Saas J, Ziegelbauer K, von Haeseler A, Fast B, Boshart M (2000) A developmentally 20. Coustou V, et al. (2006) Fumarate is an essential intermediary metabolite produced by regulated aconitase related to iron-regulatory protein-1 is localized in the cytoplasm the procyclic Trypanosoma brucei. J Biol Chem 281:26832–26846. and in the mitochondrion of Trypanosoma brucei. J Biol Chem 275:2745–2755. 21. Bringaud F, Baltz D, Baltz T (1998) Functional and molecular characterization of a 39. Colasante C, Ellis M, Ruppert T, Voncken F (2006) Comparative proteomics of glyco- glycosomal PPi-dependent enzyme in trypanosomatids: Pyruvate, phosphate dikinase. somes from bloodstream form and procyclic culture form Trypanosoma brucei brucei. Proc Natl Acad Sci USA 95:7963–7968. Proteomics 6:3275–3293. 22. Wirtz E, Leal S, Ochatt C, Cross GA (1999) A tightly regulated inducible expression 40. Delauw MF, et al. (1985) Inactivation and reactivation of a variant-specific antigen system for conditional gene knock-outs and dominant-negative genetics in Trypano- gene in cyclically transmitted Trypanosoma brucei. EMBO J 4:989–993. soma brucei. Mol Biochem Parasitol 99:89–101. 23. Ngo H, Tschudi C, Gull K, Ullu E (1998) Double-stranded RNA induces mRNA degrada- 41. Heape AM, Juguelin H, Boiron F, Cassagne C (1985) Improved one-dimensional tion in Trypanosoma brucei. Proc Natl Acad Sci USA 95:14687–14692. thin-layer chromatographic technique for polar lipids. J Chromatogr 322:391– 24. Bringaud F, et al. (2000) Characterization and disruption of a new Trypanosoma brucei 395. repetitive flagellum protein, using double-stranded RNA inhibition. Mol Biochem 42. Macala LJ, Yu RK, Ando S (1983) Analysis of brain lipids by high performance thin-layer Parasitol 111:283–297. chromatography and densitometry. J Lipid Res 24:1243–1250. 25. Besteiro S, et al. (2002) Succinate secreted by Trypanosoma brucei is produced by a 43. Coustou V, et al. (2005) A mitochondrial NADH-dependent fumarate reductase in- novel and unique glycosomal enzyme, NADH-dependent fumarate reductase. J Biol volved in the production of succinate excreted by procyclic Trypanosoma brucei. J Biol Chem 277:38001–38012. Chem 280:16559–16570. BIOCHEMISTRY

Riviere et al. PNAS ͉ August 4, 2009 ͉ vol. 106 ͉ no. 31 ͉ 12699 Downloaded by guest on September 28, 2021