A Mitochondrial NADH-dependent Fumarate Reductase Involvedin the Production of Succinate Excreted by ProcyclicTrypanosoma brucei*
Received for publication, January 11, 2005, and in revised form, February 9, 2005Published, JBC Papers in Press, February 17, 2005, DOI 10.1074/jbc.M500343200
Virginie Coustou‡, Sebastien Besteiro‡§, Loıc Riviere‡, Marc Biran¶, Nicolas Biteau‡,Jean-Michel Franconi¶, Michael Boshart�, Theo Baltz‡, and Frederic Bringaud‡**
From the ‡Laboratoire de Genomique Fonctionnelle des Trypanosomatides, UMR-5162 CNRS, Universite Victor SegalenBordeaux 2, 146 rue Leo Saignat, 33076 Bordeaux cedex, France, ¶Resonance Magnetique des Systemes Biologiques,UMR-5536 CNRS, Universite Victor Segalen Bordeaux 2, 146 rue Leo Saignat, 33076 Bordeaux cedex, France, and the�Department Biology I, Genetics, Ludwig Maximilian University Munich, D-80638 Munchen, Germany
Trypanosoma brucei is a parasitic protist responsiblefor sleeping sickness in humans. The procyclic stage ofT. brucei expresses a soluble NADH-dependent fumaratereductase (FRDg) in the peroxisome-like organellescalled glycosomes. This enzyme is responsible for theproduction of about 70% of the excreted succinate, themajor end product of glucose metabolism in this form ofthe parasite. Here we functionally characterize a newgene encoding FRD (FRDm1) expressed in the procyclicstage. FRDm1 is a mitochondrial protein, as evidencedby immunolocalization, fractionation of digitonin-per-meabilized cells, and expression of EGFP-taggedFRDm1 in the parasite. RNA interference was used todeplete FRDm1, FRDg, or both together. The analysis ofthe resulting mutant cell lines showed that FRDm1 isresponsible for 30% of the cellular NADH-FRD activity,which solves a long standing debate regarding the exist-ence of a mitochondrial FRD in trypanosomatids. FRDgand FRDm1 together account for the total NADH-FRDactivity in procyclics, because no activity was measuredin the double mutant lacking expression of both pro-teins. Analysis of the end products of 13C-enriched glu-cose excreted by these mutant cell lines showed thatFRDm1 contributes to the production of between 14 and44% of the succinate excreted by the wild type cells. Inaddition, depletion of one or both FRD enzymes resultsin up to 2-fold reduction of the rate of glucose consump-tion. We propose that FRDm1 is involved in the mainte-nance of the redox balance in the mitochondrion, asproposed for the ancestral soluble FRD presumablypresent in primitive anaerobic cells.
Fumarate reductases (FRDs)1 catalyze the reduction of fu-marate to succinate and can be divided into two classes of
enzymes: those belonging to a multimeric complex associatedwith the respiratory chain and transferring electrons from aquinol to fumarate and the soluble enzymes, which transferelectrons from a noncovalently bound cofactor (NADH orFADH2/FMNH2) to fumarate. Most of the FRDs characterizedso far belong to the first class (1). These FRDs are structurallysimilar to succinate dehydrogenases (SDH), complex II of therespiratory chain. The Shewanella putrefaciens FRD is func-tionally equivalent to the membrane-bound enzymes by accept-ing/transferring electrons from/to the respiratory chain but is asoluble enzyme lacking a membrane anchor (2). To date, onlytwo examples of soluble FRDs not linked to the respiratorychain (second class) have been described. The yeast Saccharo-myces cerevisiae expresses two soluble FRDs (cytosolic andpromitochondrial) that use FADH2/FMNH2 as an electron do-nor (3, 4), and the African trypanosome Trypanosoma bruceiexpresses a soluble NADH-dependent FRD located in the per-oxisome-like organelle, called glycosome (5, 6). A phylogeneticanalysis of FRDs and SDHs showed that the membrane-boundenzymes form a monophyletic group distantly related to thesoluble enzymes, including the S. putrefaciens FRD (5). In1980, Gest (7) proposed that the membrane-bound FRDsevolved from an ancestral soluble enzyme and that the newfeatures contributed to improvement of the energy-transducingsystem. In this model, the primitive organisms used anaerobiclactic fermentation of hexoses to generate energy. The metab-olism subsequently evolved to “succinic fermentation” by intro-ducing a couple of new enzymes, including a soluble FRD thatwas possibly NADH-dependent (7, 8). Then association of FRDwith the electron transfer chain in the membrane of the mito-chondrion (first class of FRD) evolved to increase the rate ofATP production. The soluble FRDs of yeast and trypanosomesmight therefore be the closest relatives of the ancestral FRDcharacterized so far.
Trypanosomatids are protozoan parasites of major medicaland veterinary significance. They cause serious diseases inhumans such as sleeping sickness (T. brucei), Chagas disease(Trypanosoma cruzi), and Leishmaniasis (Leishmania spp.)and are important pathogens for domestic animals (Trypano-soma spp.) and for plants (Phytomonas). RNA interference(RNAi), a very powerful reverse genetic tool that allows specificinhibition of the expression of a target gene, has been recentlydeveloped in T. brucei (9–14), but the pathway seems to benonfunctional in T. cruzi and Leishmania spp (15, 16). There-fore, T. brucei has become a favorite model to study metabolismand other processes shared by trypanosomatid species. T. bru-cei differentiates into several adaptative forms during its life
* This work was supported by the CNRS, the Conseil Regionald’Aquitaine, the Ministere de l’Education Nationale de la Recherche etde la Technologie, and the International Scientific Cooperation Projects(INCO-DEV) Program of the European Commission. The costs of pub-lication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Present address: Wellcome Centre for Molecular Parasitology,Anderson College, University of Glasgow, 56 Dumbarton Road, Glas-gow, G11 6NU, UK.
** To whom correspondence should be addressed. Tel.: 33-5-57-57-46-32; Fax: 33-5-57-57-48-03; E-mail: [email protected].
1 The abbreviations used are: FRD, fumarate reductase; SDH, succi-nate dehydrogenase; RNAi, RNA interference; EGFP, enhanced greenfluorescent protein; PBS, phosphate-buffered saline; ASCT, acetate:succinate CoA transferase; PPDK, pyruvate phosphate dikinase.
cycle. The bloodstream form in the mammalian host and theprocyclic form in the midgut of the tsetse fly vector can be easilykept and genetically manipulated in axenic cell culture.
The energy metabolism of the T. brucei bloodstream form isunique among trypanosomatids. This form relies entirely onglycolysis (which takes place in the glycosome) with pyruvateas the main excreted end product, whereas the mitochondrionplays a minor role (17). In contrast, the T. brucei procyclic formand the other adaptative forms of trypanosomatids have a moreelaborate energy metabolism that is characterized by aerobicfermentation of glucose, with excretion of partially oxidized endproducts, such as succinate, acetate, and lactate (17–20). Theoriginal model of the glucose metabolism in the procyclictrypanosomes, established 20 years ago (17), has recently beenrevised (21). In the updated model (Fig. 1, gray arrows), part ofthe phosphoenolpyruvate produced in the cytosol enters theglycosome to produce succinate, which is excreted (steps 13,15–17) (5). The other part of phosphoenolpyruvate is used toproduce pyruvate through pyruvate kinase in the cytosol (step
12) (5, 22). Then pyruvate enters the mitochondrion to produceacetyl-CoA and then acetate (steps 21–24) (23, 24). Moreover, itwas shown that acetyl-CoA is not metabolized in the tricarbox-ylic acid cycle (25) and that substrate level phosphorylation isessential for ATP production when glucose is the major carbonsource (22, 26).
Succinate is the main end product excreted by procyclictrypanosomes grown in SDM79 medium, which contains bothglucose and proline. Proline is stoichiometrically converted tosuccinate and carbon dioxide in the mitochondrion (25). Ap-proximately 70% of the excreted end products derived fromglucose metabolism is also succinate, most of it being producedby the glycosomal NADH-dependent FRD (FRDg) (5). However,the origin of the remaining glucose-derived succinate is un-known, and we proposed that another NADH-FRD may beinvolved (5). In the 1990s the existence of fumarate reductasein trypanosomatids was highly controversial (27–29). Turrensand co-workers identified a soluble NADH-FRD activity infractions enriched for mitochondrial markers in T. brucei and
FIG. 1. Schematic representation of the D-glucose and L-proline metabolism in the T. brucei procyclic stage. Enzymatic steps ofD-glucose and L-proline metabolism are represented by gray arrows, except the steps leading to the production of glucose-derived succinate in themitochondrion that are represented by black arrows. Excreted products (acetate, L-alanine, lactate, succinate, and CO2) are in white characters ona black background (D-glucose metabolism) or on a gray background (L-proline metabolism). The metabolic flux at each enzymatic step istentatively represented by the thickness of the arrow. The dashed arrows indicate fluxes that are supposed to be very low or not existing, underthe standard growth conditions. The enzymatic reaction leading to the production of lactate is not known and is indicated by a question mark. Theglycosomal and mitochondrial compartments and the tricarboxylic acid cycle are indicated. AA, amino acid; 1,3BPGA, 1,3-bisphosphoglycerate; C,cytochrome c; Cit, citrate; CoASH, coenzyme A; F-6-P, fructose 6-phosphate; FBP, fructose 1,6-bisphosphate; G-3-P, glyceraldehyde 3-phosphate;G-6-P, glucose 6-phosphate; GLU, glutamate; Gly-3-P, glycerol 3-phosphate; IsoCit, isocitrate; 2Ket, 2-ketoglutarate; OA, 2-oxoacid; Oxac,oxaloacetate; 3-PGA, 3-phosphoglycerate; Pi, inorganic phosphate; PPi, inorganic pyrophosphate; �SAG, glutamate �-semialdehyde; SucCoA,succinyl-CoA; UQ, ubiquinone pool. The enzymes are: 1, hexokinase: 2, glucose-6-phosphate isomerase; 3, phosphofructokinase; 4, aldolase; 5,triose-phosphate isomerase; 6, glycerol-3-phosphate dehydrogenase; 7, glycerol kinase; 8, glyceraldehyde-3-phosphate dehydrogenase; 9, phospho-glycerate kinase; 10, phosphoglycerate mutase; 11, enolase; 12, pyruvate kinase; 13, phosphoenolpyruvate carboxykinase; 14, pyruvate phosphatedikinase; 15, glycosomal malate dehydrogenase; 16, glycosomal fumarase; 17, NADH-dependent fumarate reductase (FRDg); 18, glycosomaladenylate kinase; 19, malic enzyme; 20, alanine aminotransferase; 21, pyruvate dehydrogenase complex; 22, acetate:succinate-CoA transferase; 23,unknown enzyme, possibly acetyl-CoA synthetase; 24, succinyl-CoA synthetase; 25, citrate synthase; 26, aconitase; 27, isocitrate dehydrogenase;28, 2-ketoglutarate dehydrogenase complex; 29, succinate dehydrogenase (complex II of the respiratory chain); 30, NADH-dependent fumaratereductase (FRDm1); 31, mitochondrial fumarase; 32, mitochondrial malate dehydrogenase; 33, proline dehydrogenase; 34, pyrroline-5 carboxylatedehydrogenase; 35, glutamate aminotransferase; 36, glutamate dehydrogenase; 37, glycerol-3-phosphate oxidase; 38, rotenone-insensitive NADHdehydrogenase; 39, alternative oxidase; 40, F0/F1-ATP synthase. I, II, III, and IV indicate complexes of the respiratory chain.
T. cruzi (6, 30, 31). However, it was suggested that this activitymight correspond to a reversal of the reaction catalyzed by themitochondrial SDH (27). The involvement of a putative mito-chondrial NADH-FRD in the production of succinate is stilldebated. Here, we describe two new FRD genes (FRDm1 andFRDm2) present in the genome of T. brucei that are similar tothe glycosomal FRDg gene. FRDm1 is expressed in the mito-chondrion of procyclic trypanosomes, whereas the FRDm2 geneproduct is not detectable. By using a combination of RNAi andNMR analysis of D-[1-13C]glucose metabolism, we showed thatFRDm1 is responsible for �30% of the total cellular NADH-FRD activity and for the production of approximately �25%(between 14 and 44%) of the excreted succinate from glucosemetabolism. The role of FRDm1 in the energy metabolism ofthe procyclic trypanosomes and the regulation of glucose me-tabolism will be discussed.
EXPERIMENTAL PROCEDURES
Cell Cultures—The procyclic form of T. brucei EATRO1125 andEATRO1125.T7T were cultured at 27 °C in SDM79 medium containing10% (v/v) heat-inactivated fetal calf serum and 3.5 mg�ml�1 hemin(SDM79/fetal calf serum) (32). The bloodstream form of T. brucei An-Tat1 was grown in rats and isolated by DEAE ion exchange chroma-tography, as described previously (33).
Cloning and Sequencing of the FRD Genes—The FRDg gene, whichwe previously characterized (5), was used as a �-32P-labeled probe toscreen a genomic library of T. brucei AnTat1 constructed in the c2�75cosmid vector (34). We have selected two clones, called CosFRD4 andCosFRD11, that contained a single (FRDm1) or two (FRDg and FRDm2)FRD gene copies, respectively. Both cosmid clones were mapped, andthe CosFRD11 clone DNA was HpaI- or HindIII-digested and subclonedinto the HincII- or HindIII-digested pUC18 vector (Appligene), respec-tively. The resulting recombinant colonies were screened with the�-32P-labeled FRDg probe to identify fragments containing a part or thetotality of the FRDg and FRDm2 genes. The FRDm1 gene from theCosFRD4 cosmid clone and the pUC18 clones containing FRDg orFRDm2 gene fragments were sequenced with the Big Dye Terminatorv.1.1 cycle sequencing kit (ABI PRISMTM; PerkinElmer Life Sciences),using specific oligonucleotides as primers, as described by the manu-facturer. DNA and amino acid sequences were analyzed using the DNASTRIDER program, and data base searches were done with BLAST.Multiple alignments of amino acid sequences were obtained usingMacVector 6.0.1, and the MITOPROT program (mips.gsf.de/cgi-bin/proj/medgen/mitofilter) was used to determine the putative mitochon-drial targeting signal.
Inhibition of FRDg and/or FRDm1 Genes Expression by RNAi—Theinhibition by RNAi of FRDm1 and/or FRDg expression in the procyclicforms (9) was performed by expression of stem-loop “sense/antisense”molecules of the targeted sequences (5, 10, 22) introduced in thepLew79 expression vector (kindly provided by E. Wirtz and G. Cross)(35). Construction of the pLew-FRDg-SAS plasmid, which target FRDgexpression, was described before (5). The same approach was used togenerate the pLew-FRDm1-SAS plasmid, which target the 5�-end of theFRDm1 gene (bp 1–230) and 242 bp of the untranslated region preced-ing the gene. Briefly, a PCR-amplified 520-bp fragment containing theantisense sequence (449 bp of targeted sequence plus 50 bp used as aspacer to form the loop between the annealing sense and antisensesequences) and the restriction sites added to the primers were insertedin the HindIII and BamHI restriction sites of the pLew79 plasmid. Thena PCR-amplified 496-bp fragment containing the sense sequence wasinserted, upstream of the antisense sequence, in HindIII and XhoIrestriction sites (XhoI was introduced at the 3�-extremity of the anti-sense PCR fragment). To target the FRDg and FRDm1 genes together,we produced by overlapping PCR chimerical sense and antisense mol-ecules composed of the DNA fragments used to target FRDg andFRDm1. For the chimeric sense molecule, we generated by PCR anFRDg-S fragment flanked by HindIII (5�-extremity) and the first 20 bpof the FRDm1-S fragment (3�-extremity) and an FRDm1-S fragmentflanked by the first 20 bp of the FRDg-S fragment (5�-extremity) andXhoI (3�-extremity). Then the sense chimeric fragment was produced byoverlapping PCR, using as primers the 5� and 3� primers used to PCRamplify the FRDg-S and FRDm1-S fragments, respectively, and usingas matrix the gel-purified FRDg-S and FRDm1-S fragments. The re-sulting 926-bp PCR fragment is composed of the FRDg gene/3�-untrans-lated region fragment (453 bp) followed by the FRDm1 gene/5�-untrans-
lated region fragment (449 bp) and flanked by the HindIII (5�-extremity) and XhoI (3�-extremity) restriction sites. We used the sameapproach to generate the chimeric antisense molecule, which is com-posed of the FRDm1-AS fragment containing a 50-bp extension ascompared with the FRDm1-S fragment (spacer between the sense andantisense chimeric sequences) followed by the FRDg-AS fragment (453bp) and flanked by the HindIII and XhoI (5�-extremity) and BamHI(3�-extremity) restriction sites. The HindIII/BamHI-digested chimericantisense fragment was inserted in the HindIII and BamHI restrictionsites of the pLew79 plasmid. Then the HindIII/XhoI-digested chimericsense fragment was inserted, upstream of the chimeric antisense se-quence, in the HindIII and XhoI restriction sites of the recombinantpLew79 plasmid.
The resulting plasmids (pLew-FRDg-SAS, pLew-FRDm1-SAS, andpLew-FRDg/m1-SAS containing the a chimeric construct), each have asense and antisense version of the targeted gene fragment(s), separatedby a 36- or 50-bp fragment, under the control of the PARP promoterlinked to a prokaryotic tetracycline operator. The EATRO1125 procyclicform cell line (EATRO1125.T7T), constitutively expressing the T7 RNApolymerase gene and the tetracycline repressor under the dependenceof a T7 RNA polymerase promoter for inducible control by tetracycline(10), was transformed with the plasmids. Trypanosome transfectionand selection of phleomycin-resistant clones were performed as previ-ously reported (10, 36).
Expression of EGFP-tagged FRDm1 in Trypanosomes—The pLew79expression vector was used to produce recombinant FRDm1 proteinstagged with the enhanced green fluorescent protein (EGFP) optimizedfor fluorescence and expression in mammalian cells. The EGFP gene,preceded by a multicloning site (HindIII, BclI, XhoI, AgeI, HpaI, MluI,XbaI, and NdeI) was subcloned into the HindIII and BamHI restrictionsite of the pLew79 vector. Then PCR fragments corresponding to thefirst 800 amino acids of FRDm1 (pLew-FRDm1–1/800.EGFP) or thesame region deleted for the N-terminal 68 residues (pLew-FRDm1–69/800.EGFP) were inserted into the HpaI and XbaI restriction sites of thepLew-EGFP plasmid. Both plasmids were introduced in theEATRO1125.T7T cell line, as described above.
Production of FRD Antibodies—To produce antibodies that wouldrecognize all three FRD isoforms, the fumarate reductase domain ofFRDm1 (from amino acid positions 459 to 979), which presents 72%identity with the corresponding domain of FRDg and FRDm2, wasPCR-amplified (1572 bp DNA fragments) and cloned into the pET23aexpression vector (Novagen). To raise anti-FRDm1 specific antibodies, aPCR-amplified DNA fragment (1362 bp) corresponding to the N-termi-nal domain of FRDm1 (from amino acid positions 1 to 450) was clonedin the same expression vector. This was also done for the equivalentdomain of FRDg (from amino acid positions 1 to 355) to be used forimmunoadsorption (see below). The pET23a vector, which contains sixhistidine codons between the multicloning site and a stop codon, isdesigned to express C-terminal histidine-tagged recombinant proteins,such as FRD-his (pET23a-FRD; FRD domain of FRDm1), FRDm1-his(pET23a-FRDm1; N-terminal domain of FRDm1), or FRDg-his(pET23a-FRDg; N-terminal domain of FRDg). Induction of Escherichiacoli BL21 cells, transformed with pET23a recombinant plasmids, wasperformed for 2 h at 37 °C with 1 mM isopropyl �-D-thiogalactopyrano-side. The cells were harvested by centrifugation, and the recombinantproteins were purified by nickel chelation chromatography (Novagen)according to the manufacturer’s instructions. Antisera were raised inrabbits by three injections at 15-day intervals of 200 �g of FRD-his orFRDm1-his recombinant nickel-purified proteins, electroeluted afterseparation on SDS-PAGE, and emulsified with complete (first injection)or incomplete Freund’s adjuvant. The specificity of the produced im-mune sera was assayed against the different FRD recombinant proteinsexpressed in E. coli, by Western blot analysis. Because the anti-FRDm1immune serum cross-reacted with FRDg, immunoadsorptions were per-formed by incubating 1:5 PBS diluted rabbit anti-FRDm1 for 2 h atroom temperature with the FRDg-his recombinant protein linked ontonickel beads. FRDm1-specific antibodies depleted from FRDg-cross-reacting epitope reactivity were recovered from the supernatant aftercentrifugation. The anti-FRDg immune serum previously producedagainst a synthetic peptide, corresponding to the 23 C-terminal aminoacids of FRDg, recognizes a single 120-kDa band and is considered to bespecific to FRDg (5).
Western Blot Analysis—Total protein extracts of the wild type ormutant procyclic form or bloodstream form of T. brucei (107 cells) wereseparated by SDS-PAGE (8%) and immunoblotted on Immobilon-P fil-ters (Millipore) (37). Immunodetection was performed as described (37,38) using as primary antibodies the monoclonal mouse anti-PPDK(H112) undiluted (36), the monoclonal anti-tubulin (diluted 1:100)
(Sigma), the rabbit anti-FRDg (diluted 1:100) (5), the rabbit immuno-adsorbed anti-FRDm1 (diluted 1:50), or the rabbit anti-FRD (diluted1:100) and as secondary antibodies anti-mouse or anti-rabbit IgG con-jugated to horseradish peroxidase (Bio-Rad), respectively.
Immunofluorescence Analyses—Log phase cells were fixed withformaldehyde as described before (36). The slides were incubatedwith rabbit immunoadsorbed anti-FRDm1 antibodies (diluted 1:200)and monoclonal mouse anti-hsp60 (undiluted) (39) or anti-TRYP2(undiluted) (40) followed by fluorescein isothiocyanate-conjugatedgoat anti-rabbit secondary antibody diluted 1:100 (Bio-Rad) and Al-exa Fluor 568-conjugated goat anti-mouse secondary antibody diluted1:100 (Molecular Probes), respectively. The cells were viewed with aZeiss UV microscope, and the images were captured by a MicroMax-1300Y/HS camera (Princeton Instruments) and MetaView software(Universal Imaging Corporation) and merged in Adobe Photoshop ona Macintosh iMac computer.
Digitonin Permeabilization—Procyclic EATRO1125 cells werewashed in PBS buffer and resuspended at 6.5 � 108 cells (3.3 mg ofprotein)�ml�1 in STE buffer supplemented with 150 mM NaCl and theCompleteTM Mini EDTA-free protease inhibitor mixture (Roche AppliedScience). Cell aliquots (300 �l) were incubated for 4 min at 25 °C withincreasing concentrations of digitonin, before being centrifuged at14,000 � g for 2 min, as described before (41).
Enzymatic Assays—Sonicated (5 s at 4 °C) crude extracts oftrypanosomes resuspended in cold hypotonic buffer (10 mM potassiumphosphate, pH 7.8) were tested for enzymatic activities. NADH-de-pendent FRD and glycerol-3-phosphate dehydrogenase activitieswere measured at 340 nm via oxidation of NADH, according to pub-lished procedures (5).
D-Glucose Measurements in Growth Medium—To determine the rateof D-glucose consumption, 109 procyclic cells (exponential phase) werecollected by centrifugation, washed in PBS buffer, and resuspendedin 10 ml of fresh SDM79 medium containing 10% fetal calf serum. Thecells were incubated for 8 h, and samples (50 �l) were collected at30-min intervals. Each sample was centrifuged at 14,000 � g for 30 s,and the quantity of D-glucose remaining in the supernatant wasdetermined using the Glucose Trinder (Sigma) or Glucose GOD-PAP(Biolabo SA) kits.
NMR Experiments—4 � 109 T. brucei procyclic cells were collected bycentrifugation at 1,400 � g for 10 min, washed once in PBS buffer, andincubated in 10 ml of incubation buffer (PBS buffer supplemented with24 mM NaHCO3, pH 7.3) containing 110 �mol D-[1-13C]glucose (11 mM)for 90–180 min at 27 °C. The D-glucose concentration in the mediumwas determined with the Glucose Trinder (Sigma) or Glucose GOD-PAP(Biolabo SA) kits. The integrity of the cells during the incubation waschecked by microscopic observation. After centrifugation for 10 min at1,400 � g, the supernatant was lyophilized and redissolved in 485 �l ofD2O, and 15 �l of pure dioxane was added as an external reference. 13CNMR spectra were collected at 125.77 MHz with a Bruker DPX500spectrometer equipped with a 5-mm broad band probehead. Measure-ments were recorded at 25 °C under a bi-level broad band-gated protondecoupling with D2O lock. Acquisition conditions were: 90° flip angle,22,150 Hz spectral width, 64K memory size, and 21.5 s total recycletime. The measurements were performed overnight with 2,048 scans.The spectra were obtained after a 1-Hz exponential line broadening.The specific 13C enrichment of lactate (C3), acetate (C2), and succi-nate (C2 and C3) was determined from 1H-observed/13C-edited NMR(1H/13C NMR) spectra acquired under 13C-decoupling (42), accordingto a modification of the method used by Rothman et al. (43), whereinthe detection of protons bound to 13C is based on the spin-spincoupling between directly bound 1H and 13C nuclei (JCH � 127 Hz).Two spectra were recorded from each sample; the first correspondedto protons bound to 12C and 13C carbons (spin echo spectrum), and thesecond corresponded to protons bound to 13C carbons (13C-editedNMR spectrum). Flip angles for rectangular pulses were carefullycalibrated on both radiofrequency channels before each experiment.The relaxation delay was 8 s for a nearly complete longitudinalrelaxation. The fractional 13C enrichment at selected metabolite car-bon positions was calculated as the ratio of the area of a givenresonance in the 13C-edited NMR spectrum to its area in the standardspin-echo spectrum. The reproducibility and accuracy of the methodwere assessed using several mixtures of 13C-labeled amino acids andlactate with known fractional enrichments; the relative errors in the13C enrichment determinations were �5%.2
RESULTS
Three FRD Genes in T. brucei—We previously reportedFRDg, a gene coding for a 120-kDa glycosomal protein withNADH-dependent fumarate reductase activity expressed in theprocyclic life cycle stage of T. brucei (5). A Southern blot anal-ysis with several FRD probes indicated that the T. bruceigenome contains two additional FRD-like sequences (data notshown) that were subsequently cloned from a genomic cosmidlibrary screened with the FRDg probe. Cosmid clone CosFRD4contained the new gene designated FRDm1, and CosFRD11contained the new gene designated FRDm2 in addition to thepreviously characterized FRDg gene (Fig. 2A). A BLAST searchof the T. brucei data base (www.genedb.org/) did not revealadditional FRD-related genes.
FRDm1 and FRDm2 (1232 and 877 amino acids, respec-tively) are 52.3% identical and present 56.2 and 67.5% aminoacid identity with FRDg (1142 amino acids), respectively. Aspreviously reported for FRDg (5), FRDm1 is a putative multi-functional protein composed of three different domains. TheN-terminal domain is homologous to the ApbE protein involvedin thiamine biosynthesis, the central domain is homologous tofumarate reductases, and the C-terminal domain is homolo-gous to cytochrome b5 reductases. FRDm2 only contains thecentral and C-terminal domains (Fig. 2). Interestingly, thefumarate reductase domain is the most conserved with 99%identity at the amino acid level between FRDg and FRDm2 and71.7% between FRDm1 and either FRDg or FRDm2. The first25 and 107 residues of FRDm1 and FRDm2, respectively, werepredicted to be mitochondrial-targeting sequences by theMitoProt II 1.0a4 program (44), suggesting that these two FRDisoforms are mitochondrial proteins.
Expression of FRD Isoforms—To study the expression of thethree FRD isoforms, specific antibodies were raised against theC-terminal 23 residues of FRDg (anti-FRDg) (5), the N termi-nus of FRDm1 (anti-FRDm1), the conserved central domain ofFRDm1 (anti-FRD), and the N-terminal 144 residues ofFRDm2 (anti-FRDm2). By Western blot analysis with the anti-FRD immune serum, two proteins (120 and 130 kDa) areclearly detected in procyclic trypanosomes (Fig. 3). As previ-ously reported, the 120-kDa protein is also recognized by thespecific anti-FRDg antibodies (Ref. 5 and Fig. 3), whereas the130-kDa protein is specifically recognized by the anti-FRDm1antibodies (Fig. 3). Thus, the 120- and 130-kDa proteins corre-spond to FRDg and FRDm1, respectively, and FRDg seems tobe more abundant than FRDm1 in procyclic cells. We alsonoticed a very weak signal with the anti-FRD serum in therange of 90 kDa, which might correspond to the FRDm2 iso-form (Fig. 3). However, several lines of evidences (see “Discus-sion”) indicate that FRDm2 is not expressed and that this weakand poorly reproducible signal is not related to FRDm2 andmore likely corresponds to a degradation product of eitherFRDg or FRDm1. In the bloodstream form sample, very lowintensity bands comigrating with the FRDg (120 kDa) andFRDm1 (130 kDa) bands in the procyclic sample suggest weakexpression. However, the nature of the doublet comigratingwith FRDg in the bloodstream form sample is still unclear andis under investigation.
FRDm1 Is a Mitochondrial Protein—The subcellular local-ization of FRDm1 was studied by immunofluorescence with thespecific anti-FRDm1 immune serum (Fig. 4). A “mitochondrion-like” tubular pattern is observed, compatible with the singlereticulated mitochondrion spanning the whole cell body oftrypanosomes (45). The signal obtained with anti-FRDm1 co-localizes exactly with signals obtained with monoclonal anti-bodies specific for the heat shock protein hsp60 (Fig. 4A) ortryparedoxin peroxidase (TRYP2) (Fig. 4B), which are bona fide
2 M. Biran, G. Raffard, B. Quesson, M. Merle, P. Canioni, J. M.Franconi, and E. Thiaudiere, unpublished data.
FIG. 2. Comparison of the T. brucei FRD genes. A is a schematic representation of the three FRD genes identified in the T. brucei genome,i.e. FRDm1 (accession number AY880988) and the FRDg/FRDm2 cluster (accession number AF457132 and AY880989) that were sequenced fromthe cosmid clones CosFRD4 and CosFRD11. The white, black, and gray boxes represent the ApbE, fumarate reductase and cytochrome b5 reductasedomains, respectively, and the putative mitochondrial targeting motif present at the N terminus of FRDm1 and FRDm2 is indicated by a hatchedbox. The regions targeted for RNAi are indicated by dashed lines, and the protein segments used to produce immune sera are indicated by thicklines. B shows the amino acid alignment of the three FRD isoforms. Gray boxes define identical residues. Gaps (-) were introduced to optimize thealignment. The conserved fumarate reductase domain is boxed, as well as the putative mitochondrial targeting motifs of FRDm1 and FRDm2 andthe glycosomal targeting motif (SKI) located at the C-terminal extremity of FRDg. The amino acid positions are indicated on the right margin.
mitochondrial proteins (39, 40). The mitochondrial localizationof FRDm1 was confirmed by a cellular fractionation experi-ment wherein the different membranes of the procyclictrypanosomes were differentially permeabilized by increasingconcentrations of the detergent digitonin. A Western blot anal-ysis (Fig. 5) of the soluble fractions indicates that FRDm1 andmitochondrial markers, hsp60 (39), TRYP2 (40) and acetate:succinate CoA transferase (ASCT) (24) are released together at400 �g of digitonin/mg of protein, whereas the cytosolic fuma-rase FH23 and the glycosomal markers PPDK and FRDg (5, 36)are released at much lower digitonin concentrations (60 and250 �g of digitonin/mg of protein).
To further confirm the mitochondrial localization of FRDm1,we expressed a fusion protein composed of the first 800 aminoacids of FRDm1 followed by the EGFP in procyclic trypano-somes. This EGFP-tagged FRDm1, detected by direct fluores-cence analysis, shows a “mitochondrial-like” pattern that colo-calizes with the mitochondrial hsp60 marker (Fig. 4C).Moreover, a truncated form of the EGFP-tagged FRDm1 re-combinant protein that lacks the 68 N-terminal residues relo-calized to the cytosol of the procyclic trypanosomes (Fig. 4D).These results indicate that the predicted N-terminal signalmotif of FRDm1 is required for targeting to the mitochondrion.
FRDm1 Encodes a NADH-dependent Fumarate Reduc-tase—We have previously reported that inactivation of FRDgexpression by RNAi caused a �70% reduction of the NADH-dependent fumarate reductase activity in the mutant cell linesas compared with the wild type cell line. Moreover, this activitywas not detectable in glycosomes purified from the mutant (5).We proposed that FRD encoded by (an) additional gene(s)might account for the remaining 30% NADH-dependent fuma-rate reductase activity. Therefore, we generated mutant procy-clic cell lines for RNAi-mediated repression of FRDm1 alone orof both FRDm1 and FRDg (9). The pLew79 vector (35) was usedto produce the hairpin double-stranded RNA molecules, as
described before (5, 10, 22). The recombinant pLew79-derivedplasmids were introduced into the EATRO1125.T7T cell lineexpressing the tetracycline repressor (10, 35). We have selectedfive clonal cell lines, designated �FRDm1-B4, �FRDm1-G7,�FRDg/m1-B5, �FRDg/m1-C6, and �FRDg-C2, the latter onebeing a �FRDg clone different from the one previously ana-lyzed (5). In the absence of tetracycline, the expression of thehairpin double-stranded RNA is inhibited by the tetracyclinerepressor, which binds to the tetracyclin operator located be-tween the transcription promoter and RNAi cassette. Never-theless, all five uninduced cell lines analyzed show 9–40%reduction of the NADH-FRD activity (Fig. 6 and Table I),probably because of a leakage of the inducible expression sys-tem, as previously observed (5, 10, 22, 24, 35). After 7 days oftetracycline induction, the proteins encoded by the RNAi-re-pressed gene(s) were not detectable anymore by Western blotanalysis, for all five cell lines analyzed (Fig. 6). PPDK expres-sion, serving as specificity control, was not affected. After 7days of induction, the �FRDg cell line (�FRDg-C2.i) showed a69% reduction of the NADH-FRD activity compared with theuntransfected EATR01125.T7T cell line (22 � 3 versus 70 � 4.9milliunits�mg�1 of protein) (Table I and Fig. 6A). Upon repres-sion of FRDm1, the NADH-FRD activity was reduced by 30%(49 � 5.5 versus 70 � 4.9 milliunits�mg�1 of protein) and 26%(52 � 4.9 versus 70 � 4.9 milliunits�mg�1 of protein) for the�FRDm1-G7.i and �FRDm1-B4.i clones, respectively. This in-dicated that FRDg and FRDm1 together account for most of thecellular NADH-FRD activity (Table I and Fig. 6, A–C). This wasconfirmed by the analysis of the two double mutant clones(�FRDg/m1-B5.i and �FRDg/m1-C6.i), which displayed nomeasurable FRD activity, after 7 days of tetracycline induction(Table I and Fig. 6, D and E).
Effect of FRDm1 Depletion on Survival of Procyclic FormCells—As we described previously for the induced �FRDg celllines (5), the growth of the induced �FRDm1 and �FRDg/m1clones was not significantly affected (doubling time between12 h and 14 h compared with 13.8 h for EATRO1125.T7T).However, in some experiments initial repression of FRDm1(e.g. in clone �FRDg/m1-C6.i), but not repression of FRDg,resulted in significant cell death. No further cell mortalityoccurred after 2–3 days under induction, suggesting an adap-tative compensation of reduced FRDm1 expression. This im-plies that FRDm1 expression is important for the procyclicmetabolism. This interpretation is strengthened by our inabil-ity to obtain �FRDm1 and �FRDg/m1 mutant cell lines consti-tutively producing double-stranded RNA (data not shown),whereas such constitutive RNAi expression clones were easilyobtained for FRDg as target (5). The leakiness of the inducibleexpression system observed in all the analyzed clones mightpreadapt the parasites to further depletion of FRDm1 inducedby tetracycline.
Effect of FRD Activity Depletion on Glucose Catabolism—Therate of glucose consumption of the parental EATRO1125.T7Tline and induced �FRDg, �FRDm1-B4 and �FRDg/m1-B5clones was measured during 8 h in SDM79 medium, as de-scribed previously (5, 24). All of the mutant cell lines show areduced rate of glucose consumption, as compared with the wildtype cells (Fig. 7). The rate is the lowest in the absence of bothFRDm1 and FRDg expression (0.77 � 0.04 �mol�h�1�mg�1 ofprotein). For the induced �FRDg and �FRDm1-B4 clones, glu-cose consumption rates of 0.99 � 0.04 and 1.37 � 0.05�mol�h�1�mg�1 of protein were measured, respectively. This cor-responds to 54, 40, and 17% reduction compared with the un-transfected EATRO1125.T7T cells (1.66 � 0.06 �mol�h�1�mg�1
of protein).To further explore the role of the different FRD proteins in3 V. Coustou, S. Besteiro, and F. Bringaud, unpublished data.
FIG. 3. Expression of the FRD isoforms. Lysates (2 � 107 cells) ofT. brucei EATRO1125 procyclic form (PF) and AnTat1 bloodstreamform (BF) were analyzed by Western blotting with the immune seraindicated below each blot. The same samples were analyzed with anti-FRD and anti-tubulin (tubulin serving as a constitutively expressedcontrol in T. brucei). The procyclic samples were also analyzed withisoform-specific anti-FRDg and anti-FRDm1 immune sera. The posi-tions of the molecular mass markers are indicated in kDa.
glucose catabolism, 13C NMR spectroscopy was used to com-pare the metabolic end products excreted by untransfected andmutant (�FRDg, �FRDm1 and �FRDg/m1) cell lines. The par-asites were incubated in PBS/NaHCO3 medium containing 110�mol of D-[1-13C]glucose as the only carbon source. After ap-proximately half of D-glucose was consumed, the 13C-enrichedmolecules excreted in the incubation medium were analyzed,and the results are shown on Table I. The EATRO1125.T7Tprocyclic cell line mainly excretes succinate, acetate, and lac-tate (67.2, 20.3, and 7.9% of the excreted 13C-enriched mole-cules, respectively) and small amounts of malate, fumarate andalanine (2.9, 0.6, and 1.1%, respectively), with a total of 699 �160 nmol of 13C-enriched excreted molecules�h�1�mg�1 of cel-lular proteins (Table I and Fig. 8). For both �FRDg/m1 clones(B5 and C6 in Table I) NMR analyses were performed atdifferent time points after induction to compare different levelsof repression. These analyses led to six main observations.First, the rate of excretion of 13C-enriched molecules is 1.4–1.7times reduced in induced mutant cell lines, as compared withthe wild type or noninduced cell lines (Table I), which is inagreement with the observed reduction of the D-glucose con-sumption rate (Fig. 7). Second, significant correlations wereseen between the level of NADH-FRD activity and the relativeamounts of excreted succinate, malate, and fumarate (Fig. 9).The quantitative correlation between the NADH-FRD activityand the amount of excreted succinate, the product of FRD,confirmed that FRDm1 and FRDg confer NADH-FRD activityin vivo. This is also supported by the inverse correlation be-tween NADH-FRD activity and fumarate (substrate of FRD)and malate excretion. The accumulation of malate is probably
due to fumarase activity, which reversibly converts the accu-mulated fumarate into malate (steps 16 and 31 in Fig. 1). Third,inhibition of FRDm1 expression induces a 7% (induced�FRDm1-B4 clone) and 14% (induced �FRDm1-G7 clone) re-duction of succinate excretion, as compared with the parentalEATRO1125.T7T cell line (62.3 and 58% versus 67.2% of the13C-enriched excreted molecules, respectively), showing thatFDRm1 produces part of the excreted succinate. Fourth, in theabsence of any detectable NADH-FRD activity (induced�FRDg/m1-B5 and �FRDg/m1-C6 clones), the relative amountof excreted succinate is reduced by 93 and 91.5%, respectively,as compared with the parental EATRO1125.T7T cell line (4.7and 5.7% versus 67.2% of the 13C-enriched excreted molecules).This indicated that most, if not all, of the succinate excretedfrom glucose metabolism is produced by the glycosomal andmitochondrial FRDs. Fifth, reduction of the NADH-FRD activ-ity (FRDg and/or FRDm1) is associated with a significant re-duction of acetate production, with a maximum of 75% reduc-tion observed for the induced �FRDg/m1-B5 clone comparedwith wild type cells (4.9% versus 20.3% of the 13C-enrichedexcreted molecules) (Table I and Fig. 9). Acetate is producedfrom acetyl-CoA by ASCT, which transfers the CoA to succi-nate, to produce succinyl-CoA and by at least one additionaluncharacterized reaction (23, 24) (steps 22 and 23 in Fig. 1).Interestingly, there is a direct correlation between reduction ofacetate and succinate excretions (Fig. 9), suggesting that a highintramitochondrial succinate concentration is required to sup-port acetate production by the ASCT, as previously proposed(23). Sixth, all the mutant cell lines show a significant produc-tion of 13C-enriched glycerol, whereas this end product is not
FIG. 4. Immunolocalization ofFRDm1. Wild type procyclic T. brucei (Aand B) and transgenic procyclic cell linesoverexpressing EGFP-tagged FRDm1 (Cand D) were labeled with immunoad-sorbed (see “Experimental Procedures”)anti-FRDm1 antiserum (A and B) and amonoclonal anti-hsp60 antiserum (A, C,and D) or a monoclonal anti-TRYP2 (B).EGFP-tagged FRDm1 (amino acids1–800, C) and N-terminally truncatedEGFP-tagged FRDm1 (amino acids 69–800, D) were visualized by EGFP fluores-cence. For orientation the cells were alsostained with 4�,6�-diamino-2-phenylin-dole (DAPI) and visualized in phasecontrast.
FIG. 5. Subcellular localization ofFRDm1 by digitonin titration. Thepresence of FH2, PPDK, hsp60, TRYP2,ASCT, FRDg, and FRDm1 was deter-mined by Western blot analysis in thesupernatant from cells incubated with0–500 �g of digitonin�mg�1 of protein inSTE buffer containing 150 mM NaCl, asindicated. The concentrations of digitoninrequired to fully release proteins from cy-tosol (c), glycosomes (g), and mitochon-drion (m) are indicated by arrowheads be-low the blots.
detected in the EATRO1125.T7T supernatant. Glycerol is pro-duced in the glycosome from dihydroxyacetone phosphatethrough the formation of glycerol-3-phosphate followed by areversal of the glycerol kinase reaction (step 7 in Fig. 1), as canhappen under anaerobiosis in bloodstream forms (46, 47); noglycerol-3-phosphatase activity nor sequences that could codefor such an enzyme have been detected in T. brucei.
DISCUSSION
During the last 10 years, the existence of a soluble NADH-dependent fumarate reductase in trypanosomatids was subjectto controversy (27–29). We recently characterized a NADH-FRD (FRDg) in the glycosomes of procyclic T. brucei that isresponsible for the excretion of succinate derived from glucose
catabolism (5). However, the FRDg-depleted cell line showed aresidual NADH-FRD activity and still excreted succinate fromglucose, suggesting that procyclic cells contain at least twodifferent NADH-FRDs (5). Here, we analyzed two additionalgenes (FRDm1 and FRDm2) related to FRDg. Three lines ofevidence show that FRDm1 is a mitochondrial protein, (i) ananti-FRDm1 immune serum shows a mitochondrial immuno-fluorescence pattern, (ii) the same digitonin concentration isrequired to release FRDm1 and mitochondrial markers fromthe cells, and (iii) the first 68 N-terminal residues of FRDm1,which contain a predicted mitochondrial targeting signal, arenecessary to target the FRDm1-EGFP fusion protein to themitochondrion. FRDm1 is a bona fide NADH-FRD, because theresidual nonglycosomal NADH-FRD activity previously ob-
FIG. 6. Inactivation of FRD gene expression by RNA interference. The FRD activity and the level of FRD protein expression in single generepression clones �FRDg-C2 (A), �FRDm1-B4 (B), and �FRDm1-G7 (C) and in double gene repression clones �FRDg/m1-B5 (D) and �FRDg/m1-C6(E) were determined periodically during 7 days. FRD activity is represented as percentages of the FRD activity in the wild type cell lineEATRO1125.T7T (filled squares). Tetracycline (1 �M) was added (.i) or not (.ni) at the beginning of the experiment for each clone. The presenteddata are representative of at least three independent experiments performed for each clone. The incubation time in the presence of tetracyclinebefore the metabolic analyses shown in Table I is indicated by arrows (i). Western blots show the level of protein expression using the anti-FRDand the anti-PPDK (control) immune sera.
served in the �FRDg mutant is not detectable anymore in the�FRDg/m1 double mutant. We also demonstrated by NMRanalysis of the D-[1-13C]glucose metabolism that a significantpart of the excreted succinate is produced by FRDm1. Thisbrings additional information to our revised model of glucosemetabolism in procyclic trypanosomes (27); part of the malateproduced inside the glycosome should be transferred to themitochondrion to be converted by fumarase and FRDm1 tosuccinate (Fig. 1, black arrows). Thus, succinate excretion re-sults from a flux through part of the tricarboxylic acid cycle inthe reverse sense.
Turrens and colleagues described a NADH-FRD activity inprocyclic trypanosomes that cofractionated with other mito-chondrial enzyme activities (31); however, direct evidence forthe subcellular localization was missing. The identification ofFRDm1 as the mitochondrial NADH-FRD seems to bring thelong standing debate regarding the existence and the subcel-lular localization of FRD enzymes in trypanosomatids (27–29)to an end. NADH-FRD activity was also detected in mitochon-drial extracts of T. cruzi (30, 48) and Leishmania donovani (49).In agreement with these biochemical data, the recently se-quenced L. major and T. cruzi genomes contain one and twoFRD genes with putative mitochondrial targeting motifs, re-
spectively, in addition to a glycosomal FRD gene (data notshown). In the Trypanosoma congolense procyclic form (50) andPhytomonas sp. (51), also a NADH-FRD activity was found,although its subcellular localization is unknown. Because alltrypanosomatids analyzed to date, with the exception of thebloodstream form of T. brucei, excrete succinate under aerobicconditions (18) and express NADH-FRD activity, the proposedmodel of glucose catabolism might be extended from T. bruceito the other trypanosomatids.
Membrane-bound SDH activity (complex II of the respiratorychain; step 29 in Fig. 1), which produces fumarate from succi-nate (the reverse of the NADH-FRD reaction but with ubiqui-none as the electron acceptor instead of NAD), has beendetected in several trypanosomatids, including T. brucei (48,52, 53). The presence of both SDH and NADH-FRD activities inthe same subcellular compartment may generate a futile cycleand be detrimental for the cells. Turrens and colleaguesshowed that, in isolated mitochondrial fractions, succinate isthe main electron donor to the respiratory chain and proposeda model involving both FRD and SDH, the latter having acentral role (54). However, the analysis of SDH-depleted pro-cyclic trypanosomes does not support this hypothesis, becausethe ATP production and cell growth are not affected in the
TABLE IExcreted end products of D-[1-13C]-glucose metabolism by procyclic T. brucei cell lines
ND, not detectable.
a Number of experiments for each cell line.b Code used in Fig. 9 to identify cell lines. WT, wild type.c The mean,sd of two to five experiments are presented; the results are shown � S.D. when more than two experiments have been performed per
cell line.d Metabolic data from Ref. 5.e The FRD activity was determined on the new �FRDg-C2.i cell line.f The time of tetracycline induction, before the metabolic analysis, is shown in Fig. 6.g When only two experiments have been analyzed per cell lines, the deviation of the mean (difference between the values divided by 2) is lower
mutant cell line (26). Actually, one may consider that detecta-ble SDH activity in an in vitro assay does not necessarily implythat it functions in vivo, under the conditions studied, asclearly illustrated by the apparent absence of a carbon fluxthrough the complete tricarboxylic acid cycle in procyclictrypanosomes (25). Also, proline is converted quantitatively tosuccinate and carbon dioxide as final excreted/released endproducts (25). The absence of net production of fumarate ormalate from proline sustains the view that SDH is not impor-tant in procyclic trypanosomes under standard growthconditions (Fig. 1).
The presence in the mitochondrion of an aerobic organism ofSDH, which is a hallmark of aerobic eukaryotes, and FRD onlyfound in anaerobic organisms is rather unique and seems a likecontradiction. The procyclic trypanosomes need oxygen, whichis the only terminal electron acceptor used to maintain therespiratory chain activity. However, in the presence of glucose,these parasites behave like anaerobic organisms. Indeed, theconsumed carbon sources (mainly glucose, proline, and threo-nine) are converted into partially oxidized end products (succi-nate, acetate, lactate, and glycine) by the so called aerobicfermentation (17–20, 55), and ATP is primarily produced bysubstrate level phosphorylation (22, 26, 56). Consequently, thepresence of a FRD in the mitochondrion of these parasites isnot surprising. However, the data do not support a role for themitochondrial SDH under the conditions analyzed (glucose-rich medium), and therefore the role of this enzyme, typical ofaerobic metabolism, in procyclic trypanosomes remainsunclear.
In the absence of detectable NADH-FRD activity and detect-able FRD proteins (�FRDg/m1 cell lines), the production ofsuccinate is reduced by 92% as compared with the wild typecells. The limitations of our assays leave it open whether ad-ditional activity caused by the presence of a different fumaratereductase or residual activity as a result of incomplete RNAi-mediated repression may account for the remaining 8% ofexcreted succinate. The FRDm2 gene product is an obvious
candidate, however, we did not detect any FRDm2 gene expres-sion by Western blot analyses using specific anti-FRDm2 im-mune serum (data not shown). Also, the weak 90-kDa signalobserved with the anti-FRD immune serum is never observedin strains where the expression of FRDg and/or FRDm1 wereinhibited. FRDm2, might be expressed in different life cyclestage(s) of the parasite and/or in different growth conditions.Recently, Takashima et al. (57) clearly demonstrated that theT. cruzi dihydroorotate dehydrogenase, the fourth enzyme ofthe essential de novo pyrimidine biosynthetic pathway, usesfumarate (instead of NAD) as an electron acceptor. Becausethe equivalent pathway is present in T. brucei, at least a partof the residual succinate excreted by the double mutant may beproduced by this pyrimidine biosynthetic pathway (20).
Succinate is produced in the glycosomes and in the mitochon-drion by FRDg and FRDm1, respectively. To estimate the rel-
FIG. 7. Rate of glucose consumption by FRD-depleted celllines. D-Glucose consumption was measured during 8 h at 108
cells�ml�1 in SDM79 medium (�mol consumed per mg of protein;means � S.D. of three independent experiments) for EATRO1125.T7Twild type cells and the tetracycline induced (7 days of induction prior toassay) mutant clones. The rate of glucose consumption (�mol of D-glucose consumed�h�1�mg�1 of protein) is given as the means � S.D. ofthe slope calculated from the straight line generated for each experi-ment done with the same cell line: EATRO1125.T7T, 1.66 � 0.06;�FRDm1-B4.i, 1.37 � 0.05; �FRDg-C2.i, 0.99 � 0.04; and �FRDg/m1-B5.i, 0.77 � 0.04.
FIG. 8. 13C NMR spectra of metabolic end products excreted byprocyclic cell lines incubated with D-[1-13C]glucose. For this NMRanalysis, EATRO1125.T7T (A) or the induced double mutant clone�FRDg/m1-B5 (B) were incubated with 11 mM D-[1-13C]glucose in PBS/NaHCO3 buffer. The NMR spectra were obtained after addition of 15 �lof dioxane. Each spectrum illustrates one representative experimentfrom a set of at least three. The resonances were assigned as follows: A,acetate; Al, L-alanine; D, dioxane; F, fumarate; G, D-glucose; Gl, glyc-erol; L, lactate; M, malate; S, succinate. The position of the 13C-enrichedin each detected molecule is indicated. For succinate and fumarate C2and C3 resonances (C1 and C3 for glycerol) are undistinguishable andwere labeled 2,3 (1,3 for glycerol).
ative flux and the maximum capacity of each pathway, NMRstudies of D-[1-13C]glucose catabolism in FRD-depleted mu-tants are informative. First, in the absence of FRDg (�FRDgmutant), succinate excretion is reduced by 56% as comparedwith the wild type cells (29.8% versus 67.2% of the 13C-enrichedexcreted molecules). This implies that, under this condition, allof the excreted succinate results from the mitochondrialFRDm1 isoform, which represents 44% of the succinate ex-creted by the wild type cells. Second, in the absence of FRDm1,succinate excretion is reduced by up to 14% as compared withthe wild type cells (�FRDm1-G7.1 cell line). Similarly, we mayconclude that, in this mutant cell line, all of the excretedsuccinate results from the glycosomal FRDg isoform, whichrepresents 86% of the succinate excreted by the wild type cells.Thus, in the wild type EATRO1125.T7T cell line, FRDg andFRDm1 appear to produce between 56 and 86% and between 14and 44% of the total excreted succinate, respectively. The in-hibition of one succinate-producing pathway induces accumu-lation of fumarate and malate, which are in part converted intosuccinate by the other pathway. Consequently, the highestvalues represent the maximum capacity of the glycosomal(86%) and mitochondrial (44%) succinate producing pathways,whereas the lower values represent minimal estimations foreach of the fluxes in the wild type cells. This analysis showsthat in the wild type procyclic cells, the glycosomal flux isbetween 1.3- and 6-fold more important than the mitochondrialflux, with a mean of 2.4-fold.
We previously proposed that FRDg, which catalyzes the laststep of the glycosomal succinate producing pathway, is in-volved in the maintenance of the glycosomal redox balance byoxidizing NADH into NAD (5). Indeed, because PYK con-sumes part of the phosphoenolpyruvate produced in the cytosol
(step 12 in Fig. 1) (5, 22), both the glycosomal malate dehydro-genase and FRDg activities (steps 15 and 17 in Fig. 1) may berequired to reoxidize all the NADH produced by the glycosomalglyceraldehyde-3-phosphate dehydrogenase (step 8 in Fig. 1).In yeast, this function of maintaining the redox balance is, atgrowth under anaerobiosis, exerted by the cytosolic FRDS andmitochondrial OSM1, the other known soluble FRDs (FADH2/FMNH2-dependent enzyme) (3, 4). Interestingly, the succinateproduction pathways in trypanosomatids (and in anaerobicallygrown yeast) resembles the “succinic fermentation” that is sup-posed to be developed from the lactic fermentation by primitiveanaerobic organisms, to increase their capacity of oxidizing theNADH resulting from sugar fermentation in conjunction with adevelopment toward more metabolic complexity (e.g. availabil-ity of metabolites and free energy retrieval from the substrateby subsequent coupling of FRD to respiratory chain) (7). Wepropose that in trypanosomatids a branched version of theprimitive “succinic fermentation” evolved to provide more flex-ibility to the capacity of controlling the redox balance in boththe glycosomal and mitochondrial compartments. This viewsuggests that FRDm1 plays a role in reoxidation of the NADHproduced in the mitochondrion by the glucose and proline me-tabolism (steps 21, 28, 34, and 36 in Fig. 1).
The mitochondrion of procyclic trypanosomes contains atleast two other NADH dehydrogenases, which transfer elec-trons from NADH to the respiratory chain, one is rotenone-sensitive (complex I of the respiratory chain) and the otherrotenone-insensitive (58, 59). However, the relative contribu-tion of these different NADH-consuming enzymes in the main-tenance of the mitochondrial redox balance is unknown. In thiscontext it is interesting to note that the amount of glucose-derived succinate produced in the mitochondrion (between 14and 44% of total end product) is in the same range as theamount of glucose-derived acetate produced in the mitochon-drion (�20% of total excreted end products). The mitochon-drion produces one NADH per acetate (pyruvate dehydrogen-ase complex; step 21 in Fig. 1), whereas one molecule of NADHis consumed by FRDm1 per molecule of succinate produced.Although this observation may be incidental, the redox balancein the mitochondrion of glucose metabolism might be main-tained by an equimolar production of succinate and acetate.There is other evidence that the succinate and acetate path-ways are linked. Indeed, we observe a correlation between thecellular NADH-FRD activity and acetate production, when glu-cose is the only carbon source available (Fig. 9). Van Hellemondet al. (23) showed that ASCT, the only well characterized ace-tate producing enzymes in T. brucei (step 22 in Fig. 1), has a lowaffinity for succinate, one of its substrates. Thus, an importantreduction of succinate production, such as in the �FRDg/m1-B5.i cell line (92% reduction), may cause the observed reduc-tion of acetate production by ASCT (75% reduction in thismutant cell line). Because succinate is also the end product ofproline metabolism (25), the addition of a source of succinate(i.e. proline) in the NMR experiments (especially in the �FRDg/m1-B5.i mutant cell line) might be a means to test thishypothesis.
Viability of different RNAi repressor lines (see results) sug-gest that FRDm1 (but not FRDg) could be essential for theparasite; however, its absence can be complemented by anunknown enzyme/pathway if the cells are allowed to adapt bystepwise reduction of FRDm1. The unknown nature of thisadaptation, which may be related to the reduced rate of glucoseconsumption associated with repressed NADH-FRD activity, iscurrently under investigation. The normal growth rate of theparasites that consume less glucose suggests an alternativeenergy source. When grown in rich medium (SDM79), procyclic
FIG. 9. Correlation between FRD activity and relativeamounts of excreted end products of D-[1-13C]glucose metabo-lism. The fractions of 13C-enriched excreted molecules of succinate(dark squares), acetate (open triangles), malate (gray filled circles), andfumarate (gray open squares) are plotted as a function of relative FRDactivity at the time of the NMR experiments (taken from Table I;EATRO1125.T7T parental cell line set to 100%). The calculated corre-lation coefficients (R) is indicated on the four straight lines. All of thedata are from Table I, except the relative NADH-FRD activity and theNMR data of the �FRDg mutant (g), which come from Ref. 5. The valuesfrom clone �FRDg/m1-C6 have not been considered, because only a fewanalyses have been performed, and this cell line showed a significantcell death during the metabolic analysis, which might significantlyaffect the relative production of the excreted end products. The assign-ment of data points to individual mutant clones is indicated on top ofthe graph according to the code defined in Table I.
T. brucei use threonine, proline, glutamine, and pyruvate ascarbon sources, in addition to D-glucose (22, 55). We recentlyshowed that the rate of proline consumption increased in theabsence of D-glucose (56). The proline consumption of FRDmutants is currently under investigation.
Acknowledgments—We particularly thank D. Baltz for technicalhelp. We are grateful to P. A. M. Michels for critical reading of themanuscript.
REFERENCES
1. Van Hellemond, J. J., and Tielens, A. G. (1994) Biochem. J. 304, 321–3312. Pealing, S. L., Black, A. C., Manson, F. D., Ward, F. B., Chapman, S. K., and
Reid, G. A. (1992) Biochemistry 31, 12132–121403. Arikawa, Y., Enomoto, K., Muratsubaki, H., and Okazaki, M. (1998) FEMS
Microbiol. Lett. 165, 111–1164. Muratsubaki, H., Enomoto, K., Ichijoh, Y., Tezuka, T., and Katsume, T. (1994)
Prep. Biochem. 24, 289–2965. Besteiro, S., Biran, M., Biteau, N., Coustou, V., Baltz, T., Canioni, P., and
Bringaud, F. (2002) J. Biol. Chem. 277, 38001–380126. Denicola, A., Rubbo, H., Haden, L., and Turrens, J. F. (2002) Comp. Biochem.
Physiol. B 133, 23–277. Gest, H. (1980) FEMS Microbiol. Lett. 7, 73–778. Hannaert, V., Bringaud, F., Opperdoes, F. R., and Michels, P. A. (2003)
Kinetoplastid. Biol. Dis. 2, 1–309. Ngo, H., Tschudi, C., Gull, K., and Ullu, E. (1998) Proc. Natl. Acad. Sci.
U. S. A. 95, 14687–1469210. Bringaud, F., Robinson, D. R., Barradeau, S., Biteau, N., Baltz, D., and Baltz,
T. (2000) Mol. Biochem. Parasitol. 111, 283–29711. Bastin, P., Ellis, K., Kohl, L., and Gull, K. (2000) J. Cell Sci. 113, 3321–332812. LaCount, D. J., Bruse, S., Hill, K. L., and Donelson, J. E. (2000) Mol. Biochem.
Parasitol. 111, 67–7613. Shi, H., Djikeng, A., Mark, T., Wirtz, E., Tschudi, C., and Ullu, E. (2000) RNA
6, 1069–107614. Wang, Z., Morris, J. C., Drew, M. E., and Englund, P. T. (2000) J. Biol. Chem.
275, 40174–4017915. Robinson, K. A., and Beverley, S. M. (2003) Mol. Biochem. Parasitol. 128,
217–22816. DaRocha, W. D., Otsu, K., Teixeira, S. M., and Donelson, J. E. (2004) Mol.
Biochem. Parasitol. 133, 175–18617. Fairlamb, A. H., and Opperdoes, F. R. (1986) in Carbohydrate Metabolism in
Cultured Cells (Morgan, M. J., ed) pp. 183–224, Plenum Publishing Corp.,New York
18. Cazzulo, J. J. (1992) FASEB J. 6, 3153–316119. Tielens, A. G. M., and Van Hellemond, J. J. (1998) Parasitol. Today 14,
265–27120. Michels, P. A., Hannaert, V., and Bringaud, F. (2000) Parasitol. Today 16,
482–48921. Besteiro, S., Barrett, M. P., Riviere, L., and Bringaud, F. (2005) Trends
Parasitol., in press22. Coustou, V., Besteiro, S., Biran, M., Diolez, P., Bouchaud, V., Voisin, P.,
Michels, P. A., Canioni, P., Baltz, T., and Bringaud, F. (2003) J. Biol. Chem.278, 49625–49635
23. Van Hellemond, J. J., Opperdoes, F. R., and Tielens, A. G. (1998) Proc. Natl.Acad. Sci. U. S. A. 95, 3036–3041
24. Riviere, L., van Weelden, S. W., Glass, P., Vegh, P., Coustou, V., Biran, M., vanHellemond, J. J., Bringaud, F., Tielens, A. G., and Boshart, M. (2004)J. Biol. Chem. 279, 45337–45346
25. Van Weelden, S. W., Fast, B., Vogt, A., Van Der Meer, P., Saas, J., VanHellemond, J. J., Tielens, A. G., and Boshart, M. (2003) J. Biol. Chem. 278,
12854–1286326. Bochud-Allemann, N., and Schneider, A. (2002) J. Biol. Chem. 277,
32849–3285427. Tielens, A. G., and Van Hellemond, J. J. (1998) Biochim. Biophys. Acta 1365,
71–7828. Turrens, J. (1999) Parasitol. Today 15, 346–34829. Tielens, A. G., and Van Hellemond, J. J. (1999) Parasitol. Today 15, 347–34830. Boveris, A., Hertig, C. M., and Turrens, J. F. (1986) Mol. Biochem. Parasitol.
19, 163–16931. Mracek, J., Snyder, S. J., Chavez, U. B., and Turrens, J. F. (1991) J. Protozool.
38, 554–55832. Brun, R., and Schonenberger, M. (1979) Acta Trop. 36, 289–29233. Lanham, S. M., and Godfrey, D. G. (1970) Exp. Parasitol. 28, 521–53434. Bringaud, F., Vedrenne, C., Cuvillier, A., Parzy, D., Baltz, D., Tetaud, E., Pays,
E., Venegas, J., Merlin, G., and Baltz, T. (1998) Mol. Biochem. Parasitol. 94,249–264
35. Wirtz, E., Leal, S., Ochatt, C., and Cross, G. A. (1999) Mol. Biochem. Parasitol.99, 89–101
36. Bringaud, F., Baltz, D., and Baltz, T. (1998) Proc. Natl. Acad. Sci. U. S. A. 95,7963–7968
37. Harlow, E., and Lane, D. (eds) (1988) Antibodies: A Laboratory Manual, ColdSpring Harbor Laboratory, Cold Spring Harbor, NY
38. Sambrook, J., Fritsch, E. F., and Maniatis, T. (eds) (1989) Molecular Cloning:A Laboratory Manual., 2nd Ed., Cold Spring Harbor Laboratory, ColdSpring Harbor, NY
39. Bringaud, F., Peyruchaud, S., Baltz, D., Giroud, C., Simpson, L., and Baltz, T.(1995) Mol. Biochem. Parasitol. 74, 119–123
40. Tetaud, E., Giroud, C., Prescott, A. R., Parkin, D. W., Baltz, D., Biteau, N.,Baltz, T., and Fairlamb, A. H. (2001) Mol. Biochem. Parasitol. 116, 171–183
41. Marche, S., Michels, P. A., and Opperdoes, F. R. (2000) Mol. Biochem. Para-sitol. 106, 83–91
42. Freeman, R., Mareci, T. H., and Morris, G. A. (1981) J. Magn. Reson. 42,341–345
43. Rothman, D. L., Behar, K. L., Hetherington, H. P., den Hollander, J. A.,Bendall, M. R., Petroff, O. A., and Shulman, R. G. (1985) Proc. Natl. Acad.Sci. U. S. A. 82, 1633–1637
44. Claros, M. G., and Vincens, P. (1996) Eur. J. Biochem. 241, 779–78645. Simpson, L. (1972) Int. Rev. Cytol. 32, 139–20746. Hammond, D. J., Aman, R. A., and Wang, C. C. (1985) J. Biol. Chem. 260,
15646–1565447. Hammond, D. J., and Bowman, I. B. (1980) Mol. Biochem. Parasitol. 2, 77–9148. Christmas, P. B., and Turrens, J. F. (2000) FEMS Microbiol. Lett. 183,
225–22849. Chen, M., Bennedsen, M., Zhai, L., and Kharazmi, A. (2001) Apmis 109,
801–80850. Obungu, V. H., Kiaira, J. K., Olembo, N. K., and Njobu, M. R. (1999) Indian
J. Biochem. Biophys. 36, 305–31151. Sanchez-Moreno, M., Lasztity, D., Coppens, I., and Opperdoes, F. R. (1992)
Mol. Biochem. Parasitol. 54, 185–19952. Opperdoes, F. R., Markos, A., and Steiger, R. F. (1981) Mol. Biochem. Parasi-
tol. 4, 291–30953. Chen, M., Zhai, L., Christensen, S. B., Theander, T. G., and Kharazmi, A.
(2001) Antimicrob. Agents Chemother. 45, 2023–202954. Denicola-Seoane, A., Rubbo, H., Prodanov, E., and Turrens, J. F. (1992) Mol.
Biochem. Parasitol. 54, 43–5055. Cross, G. A., Klein, R. A., and Linstead, D. J. (1975) Parasitology 71, 311–32656. Lamour, N., Riviere, L., Coustou, V., Coombs, G. H., Barrett, M. P., and
Bringaud, F. (January 21, 2005) J. Biol. Chem. 10.1074/jbc.M41427420057. Takashima, E., Inaoka, D. K., Osanai, A., Nara, T., Odaka, M., Aoki, T., Inaka,
K., Harada, S., and Kita, K. (2002) Mol. Biochem. Parasitol. 122, 189–20058. Fang, J., Wang, Y., and Beattie, D. S. (2001) Eur. J. Biochem. 268, 3075–308259. Fang, J., and Beattie, D. S. (2002) Biochemistry 41, 3065–3072
![20060503268 - apps.dtic.mil · TGK2], and 3 each for Leishmania major [LMK1, LMK2, & LMK3] and Trypanasoma cruzi [TCK1, TCK2, & TCK3] see table). We have cloned complete cDNAs of](https://static.documents.pub/doc/80x56/5f64cbaf04c3ed0e7e4cfc0e/20060503268-appsdticmil-tgk2-and-3-each-for-leishmania-major-lmk1-lmk2.jpg)
