Phosphatidylglycerophosphate synthase associates with a mitochondrial inner membrane complex and is essential for growth of Trypanosoma brucei

Authors


For correspondence. E-mail peter.buetikofer@ibmm.unibe.ch; Tel. (+41) 31 631 4113; Fax (+41) 31 631 3737.

Summary

Maintenance of the lipid composition is important for proper function and homeostasis of the mitochondrion. In Trypanosoma brucei, the enzymes involved in the biosynthesis of the mitochondrial phospholipid, phosphatidylglycerol (PG), have not been studied experimentally. We now report the characterization of T. brucei phosphatidylglycerophosphate synthase (TbPgps), the rate-limiting enzyme in PG formation, which was identified based on its homology to other eukaryotic Pgps. Lipid quantification and metabolic labelling experiments show that TbPgps gene knock-down results in loss of PG and a reduction of another mitochondria-specific phospholipid, cardiolipin. Using immunohistochemistry and immunoblotting of digitonin-isolated mitochondria, we show that TbPgps localizes to the mitochondrion. Moreover, reduced TbPgps expression in T. brucei procyclic forms leads to alterations in mitochondrial morphology, reduction in the amounts of respiratory complexes III and IV and, ultimately, parasite death. Using native polyacrylamide gel electrophoresis we demonstrate for the first time in a eukaryotic organism that TbPgps is a component of a 720 kDa protein complex, co-migrating with T. brucei cardiolipin synthase and cytochrome c1, a protein of respiratory complex III.

Introduction

The anionic phospholipid, phosphatidylglycerol (PG), is a major constituent of bacterial membranes. In addition, it serves as precursor for a second bacterial anionic phospholipid, cardiolipin (CL). In Escherichia coli, PG and CL constitute approximately 20% and 5%, respectively, of total phospholipids (Cronan, 2003). In contrast, PG is present in trace amounts only in eukaryotic cells, where it is confined to mitochondria (Chang et al., 1998b). Because of its low abundance in eukaryotes, little is known about the structural and functional importance of PG, apart from its role as intermediate in the CL biosynthetic pathway (Schlame, 2008) and its importance as pulmonary surfactant (Gunther et al., 2001). The first committed step in PG synthesis involves the formation of phosphatidylglycerophosphate (PGP) from CDP-diacylglycerol and glycerol-3-phosphate, mediated by PGP synthase (Pgps). Following dephosphorylation of PGP by PGP phosphatase, PG is used as substrate for CL formation by CL synthase (Cls). Interestingly, this last step in CL synthesis differs between prokaryotes and eukaryotes. Most bacterial Cls use two PG molecules as substrates for CL formation whereas most eukaryotic Cls use PG and CDP-diacylglycerol to form CL (Schlame, 2008).

In E. coli, PG and CL are not strictly essential for cell viability. A Pgps null mutant (pgsAΔ) having no detectable PG and CL grows normally in nutrient-rich medium, but not in low-osmolarity or minimal media, or at elevated temperatures (Kikuchi et al., 2000). The observation that pgsAΔ cells have increased levels of phosphatidic acid compared with wild-type cells suggests that anionic phospholipids may, at least to a certain degree, functionally substitute for each other to sustain basic functions in bacteria (Matsumoto, 2001). Similarly, in Saccharomyces cerevisiae knocking out the gene encoding Pgps results in a complete loss of PG and CL (Chang et al., 1998a), with the mutant cells showing reduced growth at higher temperatures and depending on fermentable carbon sources for growth (Chang et al., 1998a). In addition, E. coli clsA knockout strains (clsAΔ) are viable, yet they grow slower and reach lower final cell densities than wild-type bacteria. Interestingly, clsAΔ strains retain low levels of CL (Shibuya et al., 1985; Nishijima et al., 1988), which was recently shown to be due to a novel cls activity in E. coli (clsC) that uses PG and PE as substrates for CL formation (Tan et al., 2012). S. cerevisiae Cls null mutants (crd1Δ) completely lack CL and grow on both fermentable and non-fermentable carbon sources, yet less efficiently on the latter (Tuller et al., 1998; Chang et al., 1998b). Although mitochondrial respiration, ATPase and cytochrome oxidase activities are reduced in yeast crd1Δ (Jiang et al., 2000; Koshkin and Greenberg, 2000), CL is not strictly required for mitochondrial function and cell growth. Furthermore, the observation that depletion of CL alone results in a less severe growth defect than depletion of both PG and CL suggests that, at least in yeast, PG may substitute for CL in certain critical mitochondrial processes. In contrast to bacteria and yeast, CL synthesis is essential for growth of Trypanosoma brucei in culture. In conditional Cls knockout cells, ablation of Cls expression results in reduced cellular CL levels, alterations in mitochondrial function, and parasite death (Serricchio and Bütikofer, 2012).

Trypanosoma brucei, a protozoan parasite causing human African sleeping sickness and nagana in domestic animals, has served as model organism in the discovery of several biological events that were later identified in other eukaryotes as well (Cross, 1977; Ferguson et al., 1988). Interestingly, members of the family of Trypanosomatidae, to which T. brucei belongs, typically contain single copies of organelles, such as the mitochondrion and the Golgi (McConville et al., 2002). Because of such a simplified intracellular organization compared with most other eukaryotes, T. brucei has been recognized as valuable model eukaryote to study organelle biosynthesis and membrane trafficking (Schneider, 2001; Panigrahi et al., 2009; Riviere et al., 2009; Choi et al., 2012).

Although T. brucei has long been known to take up and metabolize lipids from the environment (Emmermann and Schmitz, 1993; Iwata et al., 1998; Smith and Bütikofer, 2010), recent work has demonstrated that inhibition of de novo lipid biosynthesis severely affects viability of parasites in culture (Martin and Smith, 2006; Signorell et al., 2008b). In particular, downregulation of or knocking out enzymes involved in the synthesis of phospholipids has been shown to alter mitochondrial morphology and parasite growth (Signorell et al., 2009; Serricchio and Bütikofer, 2012). To further study the importance of the two anionic mitochondria-specific phospholipids, CL and PG, in T. brucei, we focused on the first committed step in PG synthesis, catalysed by Pgps. Our results show that Pgps in T. brucei is essential for parasite growth and is part of a high-molecular-mass complex in the inner mitochondrial membrane. Its downregulation by RNA interference (RNAi) blocks PG formation and destabilizes mitochondrial inner membrane protein complexes III and IV.

Results

Identification of putative Pgps in T. brucei

A T. brucei gene (Tb927.8.1720) encoding putative Pgps (TbPgps), the rate-limiting enzyme in PG synthesis, was identified by blast searches using the protein sequence of Pgs1 from S. cerevisiae (CAA09905.1) as query. Comparisons on the amino acid sequence level show 26% identity and 43% similarity between the two (predicted) proteins. Similar to S. cerevisiae Pgs1, TbPgps contains tandem phospholipase D (PLD)-like motifs, which are characteristic of enzymes belonging to a protein superfamily that includes phosphatidylserine synthases of Gram-negative bacteria, bacterial-type Cls, eukaryotic Pgps, PLDs, nucleases and pox envelope proteins (Koonin, 1996). These PLD-like enzymes catalyse transphosphatidylation reactions by the action of catalytically active histidine and lysine residues found in two conserved HxK amino acid motifs. Tb927.8.1720 encodes a protein of 615 amino acids with a calculated molecular mass of 68.1 kDa. Topology prediction using TMHMM (Krogh et al., 2001) indicates the presence of a single N-terminal transmembrane domain spanning amino acids 2–24, which may target the protein to mitochondria [according to MitoCarta predictor (Zhang et al., 2010)].

TbPgps is essential for parasite survival

To study the function of TbPgps for parasite growth, we downregulated its expression in T. brucei by expressing a tetracycline-inducible short-hairpin RNA complementary to parts of the TbPgps sequence, resulting in activation of the RNAi machinery (Zhang et al., 2011). The results show that addition of tetracycline to T. brucei procyclic forms in culture resulted in reduced growth after 72 h, followed by complete growth arrest after 96 h of induction (Fig. 1). A reduction in parasite growth was also observed when TbPgps was downregulated using RNAi in T. brucei bloodstream forms (Fig. S1). To confirm downregulation of TbPgps mRNA, Northern blot analyses after 3 days of induction of RNAi was performed (Fig. 1, insets). Together, these experiments demonstrate that expression of putative TbPgps is essential for growth of T. brucei procyclic and bloodstream forms in culture.

Figure 1.

Growth of T. brucei parasites after downregulation of TbPgps. Procyclic form RNAi parasites were cultured in the absence (squares) or presence (circles) of tetracycline. At indicated times, parasite concentrations were determined and the cultures were diluted to a density of 106 ml−1. Mean values (cumulative cell densities) ± standard deviations from three independent experiments are shown. The inset shows TbPgps mRNA analysis by Northern blotting after 3 days of incubation in the absence (−) or presence (+) of tetracycline (upper panel). To control for equal loading, rRNA was stained with ethidium bromide (lower panel).

TbPgps downregulation blocks PG synthesis and reduces PG levels

To confirm the involvement of putative TbPgps in PG formation, we analysed PG and CL levels in T. brucei procyclic forms after induction of RNAi. Quantification of phospholipids using one-dimensional thin-layer chromatography (TLC) followed by lipid phosphorus determination revealed that after 3 or 5 days of RNAi induction PG was reduced by > 80% compared with control un-induced parasites, i.e. the relative amounts of PG dropped from 1.5% to < 0.2% of total phospholipid (Fig. 2A, left panels). In addition, downregulation of TbPgps resulted in a reduction of CL by approximately 25% (Fig. 2A, right panels). No significant changes (i.e. less than 15% difference) were observed in the relative amounts of phosphatidylcholine, phosphatidylinositol plus inositol phosphorylceramide, and phosphatidylethanolamine between control and RNAi-induced parasites. The PG molecular species composition was analysed by electron-spray ionization mass spectrometry (ESI-MS), using direct infusion of total lipid extracts and acquisition of spectra in the negative ion mode, in combination with collision-induced dissociation tandem mass spectrometry (MS/MS) to identify selected molecular species. The results show a set of m/z peaks at 759.6, 773.6 and 775.6 (Fig. 2B, top panel, indicated by asterisks), which were identified as PG molecular species, with m/z 759.6 representing the ether-type species, PG(O-18:0/18:2) and PG(P-18:0/18:1), and m/z 773.6 and 775.6 representing the diacyl-type species, PG(18:2/18:0), PG(18:1/18:1) and PG(18:1/18:0) respectively. The peak assignments are in line with results from a recent T. brucei lipidomic analysis (Richmond et al., 2010). After downregulation of TbPgps, all PG molecular species were greatly reduced, or absent (Fig. 2B, bottom panel), confirming depletion of PG after ablation of TbPgps expression. In contrast, no major changes in the molecular species composition of other phospholipid classes, including CL, were observed (Fig. 2B, compare top versus bottom panel, and data not shown).

Figure 2.

Quantification of PG and CL after TbPgps RNAi.

A. TbPgps was downregulated in T. brucei procyclic forms by RNAi for 3 or 5 days by incubation in the presence of tetracycline. Subsequently, phospholipids were extracted, separated by one-dimensional TLC and individual classes quantified by lipid phosphorus determination. PG and CL levels are expressed as percentages of total phospholipids. The bars represent mean values ± standard deviations from three independent experiments.

B. ESI-MS analysis of T. brucei phospholipids before (upper panel) and after (lower panel) downregulation of TbPgps for 3 days. The m/z range between 720 and 920 is shown. Asterisks highlight the PG molecular species identified by MS/MS (see main text for details); molecular species of other phospholipid classes are indicated by horizontal lines. PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; IPC, inositol phosphorylceramide.

Direct involvement of TbPgps in PG synthesis was further studied by incubating TbPgps RNAi trypanosomes for 3 h in the presence of [3H]-labelled glycerol, which is incorporated into all major phospholipid classes, including PG and CL, in T. brucei (van Weelden et al., 2005; Serricchio and Bütikofer, 2012). Analysis of total [3H]-labelled phospholipids by one-dimensional TLC followed by radioisotope scanning revealed that after depletion of TbPgps using RNAi, production of [3H]-labelled PG was no longer detectable, whereas incorporation of radioactivity into CL was substantially reduced (Fig. 3, compare A versus B). No major changes in incorporation of radioactivity were observed for other phospholipid classes (Fig. 3). Similar results were obtained when parasites were incubated with [3H]-glycerol for longer times (6 h and 16 h; see Fig. S2). Together, these results demonstrate that RNAi-mediated downregulation of TbPgps results in inhibition of de novo synthesis of PG and depletion of cellular PG molecular species.

Figure 3.

De novo synthesis of phospholipids after depletion of TbPgps. T. brucei procyclic-form TbPgps RNAi cells were incubated for 72 h in the absence (A) or presence (B) of tetracycline to downregulate expression of TbPgps. During the last 3 h, [3H]-glycerol was added to the culture medium to label phospholipids. After extraction and separation by one-dimensional TLC, radiolabelled lipids were visualized by radioisotope scanning. The site of sample application (start) and the migration of the solvent front (front) are indicated. The migration of the major phospholipid classes is indicated. PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; IPC, inositol phosphorylceramide; CL, cardiolipin; PG, phosphatidylglycerol. The scans are representative of at least three independent experiments.

TbPgps localizes to the mitochondrion and associates with a high-molecular-mass protein complex

The subcellular localization of TbPgps in T. brucei procyclic forms was studied by expression of a C-terminally cmyc-tagged form of the protein followed by immunofluorescence microscopy. We found that cmyc-tagged TbPgps shows a similar tubular staining as the mitochondrial marker, Hsp70 (Fig. 4A). A similar staining pattern was also observed when T. brucei procyclic forms were transfected with a C-terminally GFP-tagged TbPgps (Fig. S3). To confirm mitochondrial localization biochemically, cmyc-tagged TbPgps was analysed in digitonin-isolated mitochondria using blue-native polyacrylamide gel electrophoresis (BN-PAGE) and immunoblotting (Fig. 4B). The results show that TbPgps associates with a 720 kDa protein complex (Fig. 4B, first lane). Bands of similar molecular masses were obtained when digitonin-isolated mitochondria from parasites expressing HA-tagged TbCls were probed with antibodies against TbCls and cytochrome c1 (Cyt c1), a component of the cytochrome bc1 complex (respiratory complex III) (Fig. 4B, second and third lanes), whereas staining with an antibody against Cox4, a component of the cytochrome oxidase complex (respiratory complex IV), revealed a band at 740 kDa (Fig. 4B, fourth lane). An additional band reacting with anti-cmyc (detecting TbPgps) and anti-HA (detecting TbCls) was detected at 240 kDa (Fig. 4B, first and second lanes respectively) and may represent a smaller, or partially assembled, protein complex containing TbPgps and TbCls. Together, these results show that TbPgps is localized in the mitochondrion, where it may associate with TbCls and Cyt c1 in an inner mitochondrial membrane protein complex.

Figure 4.

Localization of TbPgps by immunofluorescence microscopy and BN-PAGE analysis.

A. T. brucei procyclic forms expressing cmyc-tagged TbPgps were immunostained with antibodies against cmyc (red) and mitochondrial Hsp70 (green). DNA is shown in blue. Scale bar: 5 μm.

B. BN-PAGE and immunoblot analysis of digitonin-isolated mitochondria from T. brucei procyclic forms expressing cmyc-tagged TbPgps or HA-tagged TbCls. Native protein complexes were separated on 4–13% native gels, transferred to a membrane and detected using antibodies against Cox4, Cyt c1, HA and cmyc. Molecular mass markers are indicated in the margin.

TbPgps knock-down affects mitochondrial morphology and destabilizes inner mitochondrial membrane respiratory complexes

To determine if a reduction in PG levels affects mitochondrial integrity, parasites after downregulation of TbPgps for 96 h were stained with the mitochondrial dye, MitoTracker. The results show that the typical tubular network of the mitochondrion in control untreated T. brucei procyclic forms is altered after depletion of TbPgps (Fig. 5A). An accumulation of bright spots within the mitochondrion, which appeared less brightly stained when compared with the mitochondrion in control parasites, was observed in 50–60% of RNAi-induced parasites. However, quantification of the mitochondrial membrane potential using tetramethylrhodamine ethyl ester and flow cytometry revealed no significant decrease in the mitochondrial membrane potential after 96 h of TbPgps knock-down (Fig. S4A). In addition, no difference in the amount of reactive oxygen species produced was observed between control and TbPgps-depleted cells (Fig. S4B). In contrast, we found that downregulation of TbPgps expression affects the stability of mitochondrial protein complexes and supercomplexes (Fig. 5B). BN-PAGE and immunoblotting revealed that Cox4 and Cyt c1 progressively decreased during depletion of TbPgps (Fig. 5B, top and middle panels). In contrast, downregulation of TbPgps had no effect on the outer mitochondrial membrane porin, Vdac, which is part of a 200 kDa complex (Fig. 5B, bottom panel). Together, these data demonstrate that depletion of TbPgps affects mitochondrial integrity and mitochondrial inner membrane complex and supercomplex formation, or stability.

Figure 5.

Effect of TbPgps downregulation on mitochondrial integrity and protein complex stability.

A. T. brucei procyclic-form TbPgps RNAi cells were cultured for 96 h in the presence of tetracycline to downregulate TbPgps expression. Mitochondria were stained with MitoTracker (red) and images of live cells were acquired. Scale bar: 5 μm.

B. T. brucei procyclic-form TbPgps RNAi cells were cultured in the absence (−) or presence (+) of tetracycline (tet) to downregulate TbPgps expression. At indicated times, mitochondria were extracted using digitonin and proteins were analysed by BN-PAGE and immunoblotting. Inner mitochondrial protein complexes IV and III were visualized using antibodies against Cox4 (αCox4) and cytochrome c1 (αCyt c1), respectively, whereas an outer mitochondrial membrane complex was labelled using an antibody against Vdac (αVdac), in combination with the respective peroxidase-labelled secondary antibodies. Molecular mass markers are indicated in the margin. The blot was over-exposed to visualize the supercomplexes with apparent molecular masses > 800 kDa, containing Cox4 and cytochrome c1.

Discussion

Previous reports using E. coli and S. cerevisiae knockout mutants have shown that Pgps, the enzyme catalysing the first committed step in PG and CL synthesis, is not strictly required for cell viability (Chang et al., 1998a; Kikuchi et al., 2000). In contrast, our results demonstrate that in T. brucei procyclic and bloodstream forms TbPgps is essential for parasite growth in culture. Downregulation of TbPgps expression by RNAi resulted in a block in de novo PG synthesis and in > 80% depletion of PG levels in procyclic forms. These results establish T. brucei as the first eukaryote in which expression of Pgps is essential for cell survival, validating TbPgps as potential drug target.

The observation that incorporation of radioactivity during incubation of parasites with 3H-glycerol, indicative of de novo formation of CL, was only reduced but not completely blocked after downregulation of TbPgps can be explained in two ways. First, it is possible that residual TbPgps activity may still be present after RNAi, resulting in the formation of small amounts of PG that is immediately used for CL synthesis. However, the amounts of CL produced are not sufficient to support growth of trypanosomes, indicating that PG itself may be necessary for parasite viability. Second, the formation of 3H-labelled CL may not represent de novo CL synthesis but acyl chain remodelling. During incubation, 3H-glycerol may get converted to metabolites, such as 3H-acetate, that can be used for fatty acid synthesis or elongation (van Weelden et al., 2005; Stephens et al., 2007). These newly formed 3H-labelled fatty acids may subsequently be incorporated into CL by acyl chain remodelling, a mechanism well-known to occur in CL in mammalian cells (Hauff and Hatch, 2006). At present, it is not known if trypanosomes are capable of CL acyl chain remodelling; a homologue of one of the mammalian enzymes known to be involved in acyl chain remodelling, taffazin (Xu et al., 2003), could not be identified in the T. brucei genome. In addition, our observation that CL levels in TbPgps-depleted cells are only reduced by 25% while PG levels drop by more than 80% of control values suggests that CL may have a slower turnover than PG, i.e. its levels remain relatively stable over an extended period of time despite the lack of substrate.

In addition, we found that depletion of PG and reduction in CL in T. brucei procyclic forms results in abnormal mitochondrial morphology. These observations are in line with a recent report showing that a block in CL formation in T. brucei procyclic forms by knocking-out TbCls leads to loss of mitochondrial integrity and reduced membrane potential (Serricchio and Bütikofer, 2012). However, in contrast to TbCls-deficient parasites, knock-down of TbPgps does not result in complete disintegration of the tubular network of the mitochondrion. In addition, unlike TbCls knockout trypanosomes, TbPgps RNAi cells show no decrease in mitochondrial membrane potential and no obvious mobility phenotype. Together these results indicate that a loss of CL has a more severe effect on mitochondrial integrity, mitochondrial membrane potential and parasite mobility than depletion of PG. Interestingly, mitochondrial alterations have also been reported in T. brucei after inhibition of phosphatidylethanolamine synthesis (Signorell et al., 2009) or downregulation of mitochondrial fatty acid synthesis (Guler et al., 2008; Ralston and Hill, 2008), suggesting that the mitochondrion in T. brucei is particularly sensitive to changes in lipid composition.

Cardiolipin and PG are confined to mitochondria in eukaryotic cells. Accordingly, using subcellular fractionation, immunohistochemistry and immunofluorescence microscopy, Cls and Pgps have been localized to (inner) mitochondrial membranes in S. cerevisiae (da Cunha e Silva et al., 1989; Dzugasova et al., 1998) and mammalian cells (Furlong, 1989; Schlame and Haldar, 1993; Denny et al., 2001; Nes et al., 2012). Similarly, in a recent study functionally active, epitope-tagged TbCls was localized to the mitochondrion in T. brucei procyclic forms (Serricchio and Bütikofer, 2012). We now report using immunofluorescence microscopy and analysis of digitonin-isolated mitochondria that epitope-tagged TbPgps is also present in the T. brucei mitochondrion. In addition, BN-PAGE and immunoblotting shows that TbPgps is part of a high-molecular-mass complex and that it co-migrates with TbCls, suggesting that the two proteins may be present in the same mitochondrial membrane protein complex. Association of TbPgps and TbCls, possibly forming a functionally essential mitochondrial phospholipid synthesis complex, has previously not been demonstrated in any organism.

Cardiolipin has been shown to bind and regulate the activity and stability of several mitochondrial inner membrane complexes, including S. cerevisiae respiratory complexes III (cytochrome bc1 complex) and IV (cytochrome c oxidase complex) (Michels et al., 2000; Lange et al., 2001; Pfeiffer et al., 2003). In line with these reports, we recently showed that depletion of CL in TbCls conditional knockout trypanosomes leads to a decrease in the amounts of Cyt c1 and Cox4 (Serricchio and Bütikofer, 2012). We now demonstrate that RNAi-mediated downregulation of TbPgps expression in T. brucei procyclic forms similarly decreases Cyt c1 and Cox4 levels. It has been shown in reconstitution experiments that CL, and to a lesser extent also PG, stabilize and activate respiratory complex III in vitro (Gomez and Robinson, 1999). Together, these findings indicate that the formation and/or stability of mitochondrial respiratory complexes in T. brucei are dependent on the presence of both TbPgps and TbCls to maintain the levels of PG and CL respectively.

Experimental procedures

Unless otherwise stated, all reagents were of analytical grade and purchased from Sigma Aldrich (Buchs, Switzerland) or Merck (Darmstadt, Germany). Restriction enzymes were from Fermentas (St. Leon-Rot, Germany) and antibiotics from Sigma Aldrich, Invivogen (Nunningen, Switzerland) or Invitrogen (Basel, Switzerland). [1,2,3-3H]-glycerol (1 mCi ml−1, 60 Ci mmol−1) was purchased from American Radiolabelled Chemicals (St. Louis, MO).

Trypanosomes and culture conditions

Trypanosoma brucei 29-13 procyclic forms coexpressing a T7 RNA polymerase and a tetracycline repressor (Wirtz et al., 1999) were cultured at 27°C in SDM-79 (Invitrogen, Basel, Switzerland) containing 15% (v/v) heat-inactivated fetal bovine serum (FBS), 25 μg ml−1 hygromycin and 15 μg ml−1 G418. Bloodstream-form T. brucei derived from MiTat 1.2 (Wirtz et al., 1999) were cultured in HMI-9 (Shlomai, 2004) containing 10% (v/v) heat-inactivated FBS and 1 μg ml−1 G418. Procyclic- and bloodstream-form RNAi cells were cultured in the presence of an additional 2 μg ml−1 puromycin or 1 μg ml−1 phleomycin respectively.

RNAi-mediated gene silencing

The gene encoding putative TbPgps (Tb927.8.1720) was downregulated in T. brucei procyclic or bloodstream forms by RNAi-mediated gene silencing using a stem-loop construct containing a puromycin or phleomycin resistance gene respectively. For RNAi in procyclic forms, a 425 bp fragment of the TbPgps open reading frame (ORF) (nucleotides 993–1417) was amplified by PCR using primers Pgps_RNAi_fwd CCCAAGCTTGGATCCATTCTGTGAAGTAGCTGAGA and Pgps_RNAi_rev CCCTCTAGACTCGAGCACACTTGTATTCCTTCATT (restriction sites are underlined) and cloned into the plasmid pALC14 as described (Bochud-Allemann and Schneider, 2002), resulting in plasmid pMS1720RNAiPCF. For gene downregulation in bloodstream forms, the stem-loop construct was excised from plasmid pMS120RNAiPCF with BamHI and HindIII, and re-ligated into plasmid pLEW100v5 (http://tryps.rockefeller.edu), resulting in plasmid pMS1720RNAiBSF. Prior to transfections, plasmids were linearized with NotI. Stable clones were obtained by limiting dilution and selection with 2 μg ml−1 puromycin (procyclic forms) or 1 μg ml−1 phleomycin (bloodstream forms).

Northern blot analysis

For Northern blot analysis, total RNA was extracted from parasites using the Total SV RNA extraction Kit (Promega, Dübendorf, Switzerland) and 30 μg were loaded on a 1% agarose gel. To control for equal loading, the gel was stained with ethidium bromide after RNA separation. After transfer onto hybond-N+ nylon transfer membrane (Amersham Pharmacia Biotech) using 10× SSC buffer (150 mM Na3-citrate, pH 7.0, containing 1.5 M NaCl), the membrane was probed with a 400 bp 32P-labelled probe of the TbPgps ORF, generated using the prime-a-gene labelling system (Promega, Dübendorf, Switzerland). Hybridized probe was detected by autoradiography using BioMax MS films (GE Healthcare, Buckinghamshire, UK) in combination with intensifying screens.

Lipid analysis

Total phospholipid classes of T. brucei were extracted (Bligh and Dyer, 1959), separated by one-dimensional TLC on Silica Gel 60 plates (Merck, Darmstadt, Germany) in a solvent system composed of chloroform:methanol:acetic acid (65:25:8, v/v/v) and visualized by exposure to iodine vapour. Radioactive lipids were detected on dried TLC plates using a radioisotope detector (Berthold Technologies, Regensdorf, Switzerland) and quantified using the Rita Control software provided by the manufacturer. For quantification of lipid phosphorus, individual spots were scraped from TLC plates and quantified as described before (Signorell et al., 2008a). In each chromatographic separation, appropriate lipid standards were carried along.

Mass spectrometry

Lipids of whole cells were extracted (Bligh and Dyer, 1959), dried under N2 and stored at −20°C. Prior to analysis, lipids were dissolved in chloroform:methanol (2:1, v/v). Phospholipids were analysed using an esquire 3000 plus ion-trap mass spectrometer (Bruker Daltonics, Bremen, Germany) equipped with an API-electrospray ionization source. Samples were injected with a syringe pump and analysed in the negative ion mode using the following settings: capillary voltage: +3 kV; nebulizer pressure: 12 psi; dry gas flow rate: 5 l min−1; dry temperature: 325°C. Tandem MS/MS spectra were obtained by collision induced fragmentation using helium as collision gas. Data were collected using the Bruker Daltonics Data Analysis software provided by the manufacturer.

Construction of GFP- and cmyc-tagged TbPgps

To construct C-terminally GFP-tagged TbPgps, the Tb927.8.1720 ORF was amplified by PCR using 1720_GFP_fwd CCCAAGCTTATGCTCTTTCTTCTCGCTTACGCACTC and 1720_GFP_rev CCCTCGAGAAAAGGTCCTGCCCCATTTGTGCCAT (restriction sites are underlined). The PCR product was ligated into the HindIII- and XhoI-digested plasmid pG-EGFPΔLII β (Urwyler et al., 2007), resulting in plasmid pMS1720GFP allowing for constitutive expression of C-terminally GFP-tagged TbPgps after transient transfection into T. brucei strain 427 procyclic forms. C-terminally cmyc-tagged TbPgps was PCR-amplified with primers 1720_GFP_fwd (see above) and 1720_cmyc_rev ccCTCGAGCTAAAGGTCCTCCTCAGAAATGAGCTTCTGCTCCAAAAGGTCCTGCCCCATTTGTGC (restriction site is underlined, sequence of cmyc tag is shown in italics). The PCR product was ligated into a pLEW100-based vector (Wirtz et al., 1999). Prior to stable transfection into 29-13 procyclic forms, the plasmid was linearized with NotI.

Fluorescence microscopy

For immunofluorescence microscopy, 106 cells were allowed to adhere to a microscopy slide for 10 min. Parasites were fixed with 4% (w/v) para-formaldehyde for 10 min, washed with cold phosphate-buffered saline (PBS, 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.76 mM KH2PO4, pH 7.4), and permeabilized with 0.2% (w/v) Triton X-100 in PBS. After blocking with 2% (w/v) bovine serum albumin in PBS for 30 min, primary antibody in blocking solution was added for 45 min. Antibodies used were mouse monoclonal α-cmyc 9E10 (Santa Cruz Biotechnology, Heidelberg, Germany) and rabbit αHsp70 (kindly provided by Paul Englund, John Hopkin's School of Medicine, Baltimore) at dilutions of 1:150 and 1:1000 respectively. After washing, the corresponding secondary fluorophore-conjugated antibodies goat anti-mouse AlexaFluor568 or goat anti-rabbit AlexaFluor488 (Invitrogen, Basel, Switzerland) at dilutions of 1:100 were added in blocking solution for 45 min. After washing, cells were mounted with Vectashield containing DAPI. Fluorescence microscopy was performed with a Leica SP2 using a 100× oil objective. Pictures were acquired, processed and 3D-deconvoluted with the Leica LAS AF Version 2.1.0 software (Leica Microsystems CMS GmbH, Heerbrugg, Switzerland). For live cell staining, cells were incubated with 0.5 μM MitoTracker Red CM-H2XRos (Invitrogen, Basel, Switzerland) for 30 min, washed twice in PBS, adhered to glass slides for 5 min and immediately used for microscopy.

Quantification of mitochondrial membrane potential and reactive oxygen species

Procyclic-form parasites were treated with 125 nM tetramethylrhodamine ethyl ester perchlorate for 30 min at 27°C, washed with PBS and immediately analysed by flow cytometry. Control parasites were simultaneously treated with 50 μM carbonyl cyanide 3-chlorophenylhydrazone to cause depolarization of the inner mitochondrial membrane. Reactive oxygen species were quantified using 10 μM 2′,7′-dichlorofluorescin diacetate for 30 min at 27°C. After washing with PBS, parasites were analysed by flow cytometry.

Digitonin extraction and BN-PAGE

Mitochondria were isolated by digitonin extraction as described (Schneider et al., 2007). Briefly, 1–3 × 108 trypanosomes were washed with wash buffer (150 mM Tris-HCl, pH 7.9, 20 mM glucose, 20 mM NaH2PO4), resuspended in 0.5 ml of suspension buffer (20 mM Tris-HCl, pH 7.5, containing 600 mM sorbitol, 2 mM EDTA) followed by addition of 0.5 ml of suspension buffer supplemented with 0.05% digitonin. After incubation on ice for 5 min, non-lysed cells were removed by centrifugation at 100 g, and mitochondria were pelleted by centrifugation at 6200 g. Subsequently, mitochondria from 6.2 × 107 cells were resuspended in 100 μl of 20 mM Tris/HCl, pH 7.2, containing 600 mM sorbitol, 15 mM KH2PO4, 20 mM MgSO4 and 1.5% (w/v) digitonin, and insoluble material was removed by centrifugation for 30 min at 16 000 g. The cleared lysate was supplemented with 10× BN loading buffer [500 mM 6-aminocaproic acid, 100 mM Bis-Tris/HCl, pH 7.0, containing 5% (w/v) Coomassie Brilliant Blue G-250] and separated by BN-PAGE using a linear 4–13% (w/v) acrylamide gel as described before (Becker et al., 2008). After separation, proteins were transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA) using a semidry blot apparatus (Bio-Rad). Proteins were detected using primary antibodies against cmyc, HA (Roche, Basel, Switzerland), Cox4, Cyt c1 and Vdac (kindly provided by André Schneider, University of Bern, Switzerland) in combination with horseradish peroxidase-conjugated secondary antibodies (Invitrogen, Basel, Switzerland) and visualized with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL, USA).

Acknowledgements

We thank Monika Rauch for technical assistance during part of the study and André Schneider, Moritz Niemann and Paul Englund for antibodies and technical advice. We especially thank Manfred Heller for his support and effort during lipid mass spectrometry. The work was supported by Grant 31003A-130815 from the Swiss National Science Foundation to P.B. and, in part, by the Bern University Research Foundation. P.B. thanks A. Adkins and R. Tedder for valuable input. M.S. thanks K. Durrer for support.

Ancillary