F1000Research 2013, 2:27 Last updated: 25 DEC 2016 RESEARCH ARTICLE Proteins and lipids of glycosomal membranes from Leishmania tarentolae and Trypanosoma brucei [version 1; referees: 2 approved] Claudia Colasante1*, Frank Voncken2*, Theresa Manful3, Thomas Ruppert4, Aloysius G M Tielens5,6, Jaap J van Hellemond5, Christine Clayton4 1Institut für Anatomie und Zellbiologie, Giessen, 35392, Germany 2Department of Biological Sciences and Hull York Medical School, University of Hull, Hull, HU6 7RX, UK 3Department of Biochemistry, Cell & Molecular Biology, University of Ghana, Accra, P.O. Box LG 54, Ghana 4DKFZ-ZMBH Alliance, Zentrum für Molekulare Biologie der Universität Heidelberg, Heidelberg, D69120, Germany 5Department of Medical Microbiology and Infectious Diseases, ErasmusMC University Medical Center, Rotterdam, PO box 2040, Netherlands 6Department of Biochemistry and Cell Biology, Faculty of Veterinary Medicine, Utrecht University, Utrecht, PO Box 80176, Netherlands * Equal contributors v1 First published: 29 Jan 2013, 2:27 (doi: 10.12688/f1000research.2-27.v1) Open Peer Review Latest published: 29 Jan 2013, 2:27 (doi: 10.12688/f1000research.2-27.v1) Referee Status: Abstract In kinetoplastid protists, several metabolic pathways, including glycolysis and purine salvage, are located in glycosomes, which are microbodies that are Invited Referees evolutionarily related to peroxisomes. With the exception of some potential 1 2 transporters for fatty acids, and one member of the mitochondrial carrier protein family, proteins that transport metabolites across the glycosomal membrane version 1 have yet to be identified. We show here that the phosphatidylcholine species published report report composition of Trypanosoma brucei glycosomal membranes resembles that of 29 Jan 2013 other cellular membranes, which means that glycosomal membranes are expected to be impermeable to small hydrophilic molecules unless transport is 1 Ralf Erdmann, Ruhr-Universität Bochum facilitated by specialized membrane proteins. Further, we identified 464 proteins in a glycosomal membrane preparation from Leishmania tarentolae. Germany The proteins included approximately 40 glycosomal matrix proteins, and 2 Peter J. Myler, Seattle Biomedical homologues of peroxisomal membrane proteins - PEX11, GIM5A and GIM5B; Research Institute USA PXMP4, PEX2 and PEX16 - as well as the transporters GAT1 and GAT3. There were 27 other proteins that could not be unambiguously assigned to other compartments, and that had predicted trans-membrane domains. However, no Discuss this article clear candidates for transport of the major substrates and intermediates of Comments (0) energy metabolism were found. We suggest that, instead, these metabolites are transported via pores formed by the known glycosomal membrane proteins. F1000Research Page 1 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 Corresponding author: Christine Clayton (cclayton@zmbh.uni-heidelberg.de) How to cite this article: Colasante C, Voncken F, Manful T et al. Proteins and lipids of glycosomal membranes from Leishmania tarentolae and Trypanosoma brucei [version 1; referees: 2 approved] F1000Research 2013, 2:27 (doi: 10.12688/f1000research.2-27.v1) Copyright: © 2013 Colasante C et al. This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Data associated with the article are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication). Grant information: The work was partially supported by Deutsche Forschungsgemeinschaft, DFG112/10. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have no competing interests to declare. First published: 29 Jan 2013, 2:27 (doi: 10.12688/f1000research.2-27.v1) F1000Research Page 2 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 1. Introduction analysis of these highly purified glycosomal membrane protein In kinetoplastid protists, several metabolic pathways, including fractions did not, however, lead to the identification of any novel glycolysis, purine salvage and ether lipid biosynthesis, are located glycosomal transporters. We therefore postulate that the recently in a microbody, the glycosome1,2, which is evolutionarily related described porin activity5 in the glycosomal membrane might be pro- to peroxisomes. All evidence so far indicates that the glycosomal vided by known components of the glycosomal protein import ma- membrane, like the peroxisomal membrane, is impermeable to chinery (peroxins), as has also been suggested for peroxisomes20. nucleotides, notably adenosine phosphates and NAD(P)(H)3,4. Its In addition, we compared the phospholipid compositions of the permeability to smaller molecules, however, is subject to debate5. glycosomal membranes from bloodstream- and procyclic-form Specific transporters would be required if the membrane were im- T. brucei, with the lipid composition of the T. brucei cell membrane permeable to molecules of the size of glycolytic intermediates, such to see if this could give us more information regarding glycosomal as glucose, phosphate, malate, pyruvate, phosphoenolpyruvate and membrane transport. various triosephosphates. 2. Materials and methods In 1987, the first protein profile of glycosomal membranes from 2.1. Isolation of glycosomes from Trypanosoma brucei and Trypanosoma brucei was published6. It revealed two abundant Leishmania tarentolae proteins of 24 and 26 kDa, which were later shown to be trypano- Leishmania tarentolae promastigotes were cultured at 28°C in 3.7 some homologues of the peroxisome biogenesis protein PEX117–9. L hemin-supplemented brain-heart infusion medium to a maximum Subsequent studies, including two of the glycosomal proteome1,10, density of 2 × 108 cells/ml. Procyclic-form Trypanosoma brucei Lister revealed several more trypanosome PEX proteins that are predicted 427 was grown at 30°C in 10% (v/v) foetal calf serum-supplemented to be membrane-bound, such as PEX211,12, PEX1013, PEX1213, MEM-PROS medium to a maximum density of 5 × 106 cells/ml21. PEX1314 and PEX1415. The only transporters known to be associat- Bloodstream-form T. brucei 427 was grown at 37°C in 10% (v/v) ed with the glycosomal membrane are the ABC transporters GAT1, foetal calf serum-supplemented HMI-9 medium22 to a maximum GAT2 and GAT3, which might transport fatty acids16. In addition, density of 2 × 106 cells/ml21,23. a member of the mitochondrial carrier protein family was found: MCP6, which is a candidate for nucleotide transport17. MCP6 is Procyclic-form and bloodstream-form T. brucei (1010 cells each), found preferentially in the glycosomal membranes of bloodstream- and promastigote L. tarentolae (1012 cells) were harvested by cen- form trypanosomes, whereas in procyclic forms, it is predominantly trifugation for 10 min at 2,000× g, and were washed once in 50 ml targeted to the mitochondria17. of TEDS (25 mM Tris, 1 mM EDTA, 1 mM DTT, 250 mM sucrose, pH 7.8). After centrifugation, the cell pellet was resuspended in No analysis has yet yielded evidence for glycosomal transporters 2 ml homogenization medium (250 mM sucrose, 1 mM EDTA, of metabolites smaller than about 400 Da. In contrast, lipid bilay- 0.1% (v/v) ethanol, 5 mM MOPS, pH 7.2) containing protease ers that were reconstituted with glycosomal membrane proteins inhibitor (complete EDTA-free, Roche Applied Science) and was revealed evidence for the presence of anion- and cation-selective grinded in a pre-chilled mortar with 1 volume of wet-weight silicon pores5. The identities of these pore-forming proteins are still un- carbide (Crysalon: Norton Company: porous <400 mesh). Cells known: they could be dedicated exclusively to metabolite transport, were checked for at least 90% disruption by light microscopy. The or they might be involved in protein import as well5. cell lysate was centrifuged sequentially for 5 minutes each at 100× g and 3,000× g to remove abrasive, intact cells, cell rests and nu- If proteins other than PEX components were indeed involved in clei. The supernatant was centrifuged for 15 minutes at 17,000× g metabolite transport, it ought to be possible to find them by mass to yield the glycosome-enriched pellet fraction. This fraction was spectrometry, using highly purified glycosomal membrane protein resuspended in 3 ml of homogenization buffer and loaded on top preparations. A similar proteomics approach has been previously of a 32 ml linear 20–40% (v/v) Optiprep (iodixanol-sucrose, Sigma used for mammalian peroxisomes. Analysis of carbonate-washed rat Biochemicals) gradient, mounted on a 3.5 ml 50% (v/v) Optiprep liver peroxisomes initially yielded only two peroxisomal membrane cushion (Optiprep Application Sheet S9, Axis-shield). Centrifugation proteins, PMP70 and PMP2218, whereas a later analysis of whole was performed for 1 h at 170,000× g and 4°C using a Beckman VTi-50 mouse kidney peroxisomes led to the identification of 12 putative Rotor. 1 ml aliquots were collected from the bottom of the tube after glycosomal membrane proteins. These included one tetratricopeptide puncture, and the protein concentration of each fraction was deter- domain protein, four different ABC transporters, three members of mined using the BioRad Bradford protein assay. Of each fraction, the PMP22 family, PMP34, Pxmp4/PMP4, and the putative solute 100 µl was TCA-precipitated and the resulting pellets resuspended carrier PMP4719. in denaturing Laemmli SDS-PAGE buffer. Proteins were separated on a 12% SDS-PAGE gel and analysed by western blotting. Specific transporters for glycolytic metabolites might have been missed in previous glycosomal proteomic analyses, since glycoso- 2.2. Isolation and analysis of glycosomal membrane mal membrane proteins are likely to comprise a rather small pro- proteins from Leishmania tarentolae glycosomes portion of the total protein content. We have therefore set out to Glycosomes (corresponding to about 0.5 mg protein) were diluted identify the proteins in a highly enriched glycosomal membrane 1:5 in TEDS (see 2.1), subjected to two freeze-thaw cycles, and preparation from Leishmania tarentolae, using 30-times more centrifuged for 40 minutes at 140,000× g and 4°C. The resulting starting material than used for our previously published T. brucei pellet was washed with 5 M urea for 1 h at 4°C to remove proteins glycosomal proteome study1. Comprehensive mass spectrometry that were not tightly associated with the glycosomal membranes. Page 3 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 The glycosomal membranes were pelleted by centrifugation for 95 head-group specific mass transitions with a total dwell time of 40 minutes at 140,000× g and 4°C. This 5 M urea wash-step was 1 s, using the same settings as above. Data analysis was performed repeated once. The glycosomal membrane-enriched fraction was with Analyst™ v 1.4.1 software (MDS Sciex, Concord, ON). resuspended in denaturing Laemmli SDS-PAGE buffer and sepa- rated by SDS PAGE. In-gel trypsin digestion and nanoLC-MS/MS analysis of the obtained protein bands were performed as previ- Supplementary Table S1. F1000 Research. Claudia Colasante, ously described1. The obtained MS/MS spectra were analysed using Frank Voncken, Theresa Manful, Thomas Ruppert, Aloysius Tielens, Jaap van Hellemond, Christine Clayton Retrieved 12:12, MASCOT software and visualised in Scaffold. The comparison Jan 24, 2013 (GMT) shown in Supplementary Table 1 was done using the 2012 version of the shotgun sequence of L. tarentolae (http://tritrypdb.org/)24; only http://dx.doi.org/10.6084/m9.figshare.106948 proteins for which at least 2 different peptides could be identified with >95% confidence were included. Leishmania proteins were first scanned for the presence of a PTS1 signal based on a published 3. Results analysis for L. major and T. brucei2 and by manually examining the 3.1. Glycosome preparation from Leishmania tarentolae C-terminal sequences. Additional PTS1-containing proteins were In preliminary experiments (not shown), we tested various meth- identified using PTS1 Predictor25,26. Trans-membrane domains were ods for purification of membranes of iodixanol gradient-enriched identified using the TritrypDB annotation database. For potential T. brucei glycosomes1. The methods tested included methanol/chlo- glycosomal proteins with no known function and without an anno- roform extraction30; ultracentrifugation of glycosomes that had been tated trans-membrane domain, we also scanned for trans-membrane subjected to osmotic shock with cold water31,32; and different high domains using the HMMTOP and SOSUI algorithms27,28. salt (0–1 M NaCl or 5 M urea) washes of glycosomal membranes32. Although we were able to enrich glycosomal membrane proteins, 2.3. Phospholipid analysis as judged by the presence of PEX117, the matrix protein aldolase Lipids were extracted in triplicate from bloodstream and procyclic persisted. In addition, the total amount of membrane protein ob- T. brucei samples and from a single batch of isolated glycosomes tained from 3 × 1010 T. brucei was so low that we doubted that any due to the limited amount of purified material. Lipids were extract- lower-abundance proteins would be detected by mass spectrometry. ed according to the method of Bligh and Dyer (1959)29 with the We therefore decided to isolate glycosomes from the related kine- minor modification that 0.5% (v/v) 6 M HCl was added to the sec- toplastid L. tarentolae to increase the sensitivity for the detection ond chloroform wash to increase recovery of acidic phospholipids. of even low-abundant glycosomal membrane proteins. In contrast The phospholipids and free fatty acids were separated from neutral to T. brucei, L. tarentolae can be grown to far higher cell densi- lipids (cholesterol, cholesterol esters and triacylglycerols) by frac- ties, enabling us to isolate glycosomes from as much as 1012 cells. tionation on a 1 ml silica column prepared from 0.063–0.200 mm The different fractions obtained after differential fractionation and silica 60 (Merck, Darmstadt, Germany). Lipid extracts were dis- subsequent density gradient (Optiprep) centrifugation were ana- solved in chloroform and loaded on the silica column, then eluted lysed by western blotting (Figure 1). The gradient distributions of successively with acetone (4 volumes) and methanol (4 volumes). the two marker proteins glyceraldehyde phosphate dehydrogenase The last fraction, which contained the purified phospholipids, was (glycosomes) and HSP60 (mitochondria) are shown in Figure 1A. dried under nitrogen and stored at -20°C until HPLC-MS analysis. Comparison of previously published density gradient results from T. brucei1 with those from L. tarentolae (Figure 1A) showed that The purified phospholipids were dissolved in methanol:aceton- the mitochondria were enriched at similar gradient densities (frac- itrile:chloroform:water (46:20:17:17). Separation of molecular lipid tions 22–25), whereas the glycosomes isolated from L. tarentolae species was performed on a Synergi 4 µm MAX-RP 18A column appeared to have a higher buoyant density than those of T. brucei. In (250 × 3 mm; Phenomenex, CA, USA). Elution was performed with addition, both the mitochondria and glycosome-containing fractions a linear gradient of water in methanol/acetonitrile (60/40 (v/v)) de- were spread out over a wide range of fractions for the L. tarentolae creasing from 12.5% to 0% in 25 min, followed by further isocratic gradient, which could be the result of breakage of the organelles elution for another 25 minutes. The flow rate was kept constant at during isolation. Judging from the western blotting results (alternate 0.425 ml•min-1 and 1 µM serine and 2.5 mM ammonium acetate gradient fractions shown in Figure 1A), fractions 9, 11 and 13 con- were used in all solvents as additives. tained about 42% of the total GAPDH measured, but only 2% of the total HSP60. We therefore decided to use fractions 9–13 for further Mass spectrometry of lipids was performed using electrospray glycosomal membrane purification. ionization, on a 4000 QTRAP system (Applied Biosystems, Nieu- werkerk aan de IJssel, The Netherlands). Source temperature was To purify glycosomal membranes, we found that the protocol that set to 450°C and nitrogen was used as curtain gas. The declustering gave least matrix protein contamination was one that was successfully potential was optimized using lipid standards. The optimal colli- employed to isolate the cell membrane of E. coli33. It involved washing sion energy was dependent on the type of experiment performed the glycosomal pellet with 5 M urea, and resulted in strong depletion and was set to +45V (precursor scanning m/z 184), -45V (precur- of some prominent bands (presumably matrix proteins) and the enrich- sor scanning m/z -196), +35V (neutral loss 141), -30V (precursor ment of various proteins in the 10–25 kDa range (Figure 1B) - similar scanning m/z -241), and -40V (neutral loss scanning 87 Da) respec- to the expected sizes of the PEX11 protein homologues7,9. The entire tively. For quantification of molecular species, samples were meas- SDS-PAGE lane containing the enriched glycosomal protein fraction ured in multiple-reaction monitoring mode (MRM), monitoring for (Figure 1B) was subsequently subjected to mass spectrometry. Page 4 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 A. 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 Fraction number GAPDH HSP60 Glycosomes Mitochondria B. Glycosomal Urea Fraction pellet kDa 70 50 37 25 20 15 Figure 1. Purification of glycosomal membrane-enriched fractions from L. tarentolae. A. Western blot analysis of the different L. tarentolae fractions obtained after density gradient centrifugation. Equal volumes of only the odd-numbered fractions were loaded for analysis. Antibodies used for detection are indicated next to the western blot panels. Mitochondrial19–29 and glycosomal9–19 density gradient fractions are indicated. B. SDS-PAGE gel stained with Coomassie brilliant blue, showing protein bands from intact glycosomes (glycosomal fraction, and from the glycosomal membrane-enriched pellet (urea pellet). Arrows indicate enriched proteins in the urea-treated glycosomal membrane fraction. 3.2. Identification of putative glycosomal proteins homologues35; in addition, the L. tarentolae protein sequences in By comparison with the predicted proteome of L. tarentolae24, 464 Supplementary Table S1, Sheet 2 were manually scanned for PTS1 polypeptides were identified (Supplementary Table S1). The first signals. step that we undertook was to identify homologues of all identified proteins from the T. brucei genome (http://tritrypdb.org/tritrypdb/). 3.3. Glycosomal enzymes This was done to facilitate the retrieval of information because most The L. tarentolae glycosomal membrane preparations revealed the experimental data is available exclusively for T. brucei. All identified presence of 40 known or predicted glycosomal matrix proteins, and proteins were screened for database annotation, including user com- some novel proteins (Supplementary Table S1, sheet 2). A putative ments, and in some cases we also updated annotations from publica- glycosomal pathway scheme, incorporating all available information tions. We further screened all proteins for their presence in previously for L. tarentolae and T. brucei, is shown in Figure 2. Glycolytic en- published glycosomal1,10 and mitochondrial34 proteomes. The results zymes, enzymes involved in the conversion of glycerone phosphate are summarised in Supplementary Table S1, Sheet 1. Proteins that to glycerol, the pentose phosphate pathway, steroid and nucleotide were clearly located in compartments other than the glycosome were biosynthesis as well as enzymes of the succinic fermentation branch then excluded, resulting in Supplementary Table S1, Sheet 2. Some were detected. Similar to results obtained for the T. brucei glyco- candidates predicted to contain at least one trans-membrane do- some1, fumarase (EC 4.2.1.2) is the only enzyme of the glycosomal main were tested for their locations, by expression of N-terminally succinic fermentation branch that was not found in the L. tarentolae and/or C-terminally tagged versions (none has a PTS1 signal). The glycosomal membrane preparation. It is possible that fumarase proteins encoded by Tb927.3.1840 (putative 3-oxo-5-alpha-steroid was removed in the membrane purification; alternatively, the activ- 4-dehydrogenase), Tb927.5.1210 (putative short-chain dehydro- ity may be supplied by one of the four proteins of unknown func- genase) and Tb927.10.14020 (unknown function) were all tar- tion that are conserved in kinetoplastids and have an unambiguous geted to mitochondria, while Tb927.7.3900 (annotated as a vacu- PTS1: LtaP34.3290/Tb927.4.1360, LtaP33.2650/Tb927.11.2620, olar transporter chaperone) was in the ER (Supplementary Figure S1). LtaP18.0870/Tb927.10.13240, or LtaP24.1780/Tb927.8.6640 - al- All identified proteins were further searched for the presence of though fumarase activity would be surprising since they lack known known peroxisomal targeting signals in the L. major or T. brucei functional domains. Fumarase catalyses the conversion of malate to Page 5 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 Glycerol-3-phosphocholine Acyl-GPE Acyl-GPC 2-Ketoglutarate Isocitrate PEX2 PEX12 PEX14 PEX11 GIM5 PMP4 PEX16 ? ? ? ? ? 1.1.1.42 1.1.1.205 2.7.4.8 ? 2.7.4.3 2-Ketoglutarate Isocitrate dNTPs dNTPs IMP XMP GMP GDP ADP AMP + 2.4.2.8 ? ? 2.4.2.7 ROH + H O Tryparedoxin NADPH NADP 2.4.2.8 2.4.2.8 2 (red)2.4.2.7 2.4.2.8 NBs ? NBs Hypoxanthine Xanthine Guanine dGTP dADP Adenine Peroxidase Trypanothion ? Trypanothion 2.7.4.3 Acyl-GPE Acyl-GPC (ox) E Tryparedoxin (ox) (ox) 2.7.4.8 ROOH (ox) Arginine ? Arginine dAMP 3.1.1.5 N dGMP 2 H2O + O D 2 Ascorbate O 2.7.3.3 L Glycerol-3-phosphocholine Peroxidase Trypanothion ? Trypanothion(red) (red) (red) P-Arginine ? P-Arginine Glucono-1,5-lactone-6P 2 H2O2 MDHA DHA 5.1.3.3 1.1.1.49 O2 Glucose ? β-Glucose α-Glucose 1.15.1.1 2.4.2.10 4.1.1.23Glucosamine-6P . Orotate- Orotidine-5P UMP ? UMP 2.7.1.1 A 2.7.1.1 4 O Glycerone-3P F PRPP Mannose-1P ? Mannose-1P β-Glucose-6P α-Glucose-6P 3.1.3.37 5.4.2.8 5.3.1.9 3.5.99.6 4.1.2.13 Sedoheptulose-1,7P Sedoheptulose-7P Ribose-5P Ribose ? RiboseG 2 2.7.1.15 Mannose ? Mannose Mannose-6P β-Fructose-6P C 2.2.1.1 2.7.1.1 5.3.1.8 Erythrose-4P 2.7.1.172.7.1.11 Glyceraldehyde-3P Xylulose-5P Xylulose ? Xylulose H 2.5.1.26 3.1.3.11 Alkyl-glycerone-3P ? Alkyl-glycerone-3P Acyl-glycerone-3P β-Fructose-1,6P 2.2.1.12 2.7.1.17 2.3.1.42 4.1.2.13 Xylulose-5P Xylulose fatty acid GAT fatty acid 6.2.1.3 Glycerol ? Glycerol Glycerol-3P Glycerone-P Glyceraldehyde-3P 2-Deoxyribose-5P 2.7.1.30 1.1.1.8 5.3.1.1 4.1.2.4 acyl-CoA 1.2.1.12 2.7.1.15 I ? Glycerate-1,3P2 2-Deoxyribose HMG-CoA ? HMG-CoAtrans-enoyl-CoA 2.7.2.3 2.3.1.16 4.2.1.17 ? 1.1.1.37 ? 1.3.1.6 Glycerate-3P Oxalacetate Malate Fumarate Succinate M Mevalonateβ-hydroxyacyl-CoA 4.1.1.49 B 1.1.1.35 2.7.1.36 PEP Mevalonate-5P Mevalonate-5P2.7.9.1 ? 2.7.1.40 β-ketoacyl-CoA ? Pyruvate CoA Glycosome ? ? ? ? ? ? ? ? ? Glycerate-1,3P Glycerate-3P PEP Pyruvate CoA2 AOX Lactate Malate Fumarate Mitochondrion Mitochondrion Figure 2. Putative glycosomal pathway scheme. The scheme summarizes metabolic pathways identified so far in the glycosomes of Leishmania and T. brucei. EC numbers of enzymes identified only in the Leishmania glycosome are indicated in black boxes with white text, those identified only in T. brucei are indicated in italics, and those found in the glycosome of both species are indicated in bold. Predicted transport processes across the glycosomal membrane are indicated by the circled question marks and dashed arrows. Letters in black circles indicate the different metabolic pathways as follows: A: glycolysis; B: succinic fermentation; C: pentose phosphate pathway; D: superoxide and trypanothione metabolism; E: purine salvage; F: pyrimidine metabolism; G: mannose metabolism; H: glycerolipid biosynthesis; I: b-oxidation of fatty acids; L: phosphoarginine metabolism; M: mevalonate pathway; N: phospholipid degradation. Abbreviations used are: Acyl-GPC, 1-acyl-glycero-phosphocholine; Acyl-GPE, 1-acyl- glycero-phosphoethanolamine; AOX, alternative oxidase; DHA, dehydroascorbate; MDHA, monodehydroxyascorbate; NB, nucleobases; PRPP, 5-phosphoribosyl 1-pyrophosphate. fumarate, which is an indispensable step towards the generation of suc- the phosphomannomutase Tb927.10.6440 was found in the glyco- cinate in the glycosomal matrix. If indeed the glycosome lacks fumarase some, where it can act as phosphoglucomutase during glycolysis36. then the glycosomal membrane must harbour a malate-fumarate shut- T. brucei phospho-N-acetylglucosamine mutase (Tb927.8.980) was tle. This shuttle would be responsible for the transport of glycosome- also partially glycosomal36. Neither has an obvious PTS1 targeting derived malate in exchange with mitochondria-derived fumarate. In- signal so it is possible that they have either an internal glycosomal side the glycosome fumarate could then be converted to succinate by targeting signal or that they are co-imported via association with fumarate reductase (EC 1.3.1.6) to maintain the glycosomal NADH other glycosomal targeting signal-containing proteins37,38. The syn- balance. tenic L. tarentolae homologues LtaP36.1960 and LtaP07.0850 were not present in our dataset, but the different non-syntenic phospho- The L. tarentolae glycosomal membrane preparation contained mannomutase-like protein LtaP34.3710, containing the conserved several enzymes that were not detected during LC-MS analysis of C-terminal PTS1 signal -SKL, was found. The T. brucei aldose the T. brucei glycosome. For example, glucosamine-6-phosphate 1-epimerase, Tb927.4.1360, is the homologue of LtaP34.3290, but isomerase was found in the glycosomes of Leishmania, but not in Leishmanias have an additional isoform, LtaP35.110, containing a trypanosomes1,10. In addition, a PTS1-containing D-lactate dehy- conserved PTS1 targeting signal. drogenase-like protein is found in Leishmania, for which there is no obvious substrate, as well as a PTS1-containing xylulokinase and a The first three steps of ether-lipid biosynthesis in T. brucei, namely glucokinase-like protein. These additional enzymes involved in the the conversion of glycerone 3-phosphate to 1-alkyl-glycerol-3-phos- metabolism of sugars might indicate a higher metabolic flexibility phate, are associated with the glycosome39. T. brucei glycosome pro- of L. tarentolae compared to African trypanosomes. In T. brucei, teomic analysis confirmed the presence of the second enzyme of the Page 6 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 biosynthetic pathway, namely alkyl-glycerone-phosphate synthetase The list of known glycosomal membrane proteins that we identified (EC 2.5.1.26). The first and the third enzymes, glycerone-phosphate is shown in Table 1. It included three proteins related to PEX11 acyl-transferase (EC 2.3.1.42) and 1-acyl-glycerol-3-phosphate (PEX11, GIM5A and GIM5B), PEX2 and PEX14. The protein oxidoreductase (EC 1.1.1.101) were not identified. The T. brucei encoded by Tb09.160.4700/Tb927.9.6450 has a very weak match glycerone 3-phosphate acetyltransferase/acyltransferase homologue, to a conserved PEX16 domain (E-value 9e-3). We have therefore an- Tb927.4.3160, has a C-terminal PTS1, SRM. Analysis of the gly- notated this as a putative PEX16. It could therefore be involved cosomal membrane proteome of L. tarantolae identified not only in the incorporation of peroxisomal membrane proteins42, although alkyl-glycerone 3-phosphate synthetase (EC 2.5.1.26) but also glyc- high-throughput RNAi screening revealed no growth defects for erone 3-phosphate acetyltransferase/acyltransferase (EC 2.3.1.42, RNAi targeting this locus43. LtaP34.1280, C-terminal PTS1 -SKM) supporting the idea that the first two steps of the ether-lipid biosynthesis can occur inside the The trypanosome homologue of the mammalian peroxisomal mem- glycosome. brane protein PMP2444 (also called PXMP4 or PMP4) was also found. Like other PMP24/PMP4 proteins, a conserved TIM27 su- The b-oxidation of long chain fatty acids (LCFA) is one of the hall- perfamily domain is present in Tb927.9.1720; the function of this mark catabolic pathways attributed to peroxisomes (40 and refer- domain (and of PMP4) is still unknown. Although S. cerevisiae ences therein). Our previous analysis suggested that the glycosome of that lack it show abnormal peroxisome size and numbers45, no T. brucei was devoid of LCFA b-oxidation enzymes1. On the other growth defects were seen in the high-throughput RNAi screens in hand, 2-enoyl coenzyme A hydratase (EC 4.2.1.17) and NADP- T. brucei43. Of the previously reported T. brucei glycosomal ABC dependent 3-hydroxyacyl-CoA dehydrogenase (EC 1.1.1.35) were transporters GAT1-316, only GAT1 and GAT3 were found in the reported in glycosomal fractions from procyclic-form T. brucei40. Ac- L. tarentolae glycosomal proteome. This was not unexpected since cording to the results from this proteome analysis, the L. tarentolae gly- the annotated L. tarentolae GAT2 protein sequence is severely trun- cosome might contain all enzymes involved in the b-oxidation of LCFA cated. Either L. tarentolae lacks a functional GAT2, or this is a ge- except acyl-CoA dehydrogenase (Figure 2, Supplementary Table S1, nome assembly error; the former is quite possible since no peptides sheet 2). We found one protein (LtaP24.1780, annotated as hypo- matching L. major and L. infantum GAT2 were found. thetical), which is clearly a fatty acyl Co-A reductase and contains a C-terminal -SSL; another hypothetical protein (LtaP16.0130) contains Of the remaining potential membrane proteins, four could be tentative- -AKL and an acyl CoA dehydrogenase domain; and the enoyl-CoA ly assigned to the mitochondrion, three to the endoplasmic reticulum, hydratase/enoyl-CoA isomerase/3-hydroxyacyl-CoA dehydrogenase two to the flagellum, and one to the nucleus (Supplementary Table S1, trifunctional enzyme homologues (EC 4.2.1.17/5.3.3.8/1.1.1.35, sheet 3). The remaining proteins are listed in Table 2. A protein of the LtaP26.1590 and LtaP33.2830) containing putative PTS2 signals35. major facilitator family (Tb927.3.4070-110, LtaP29.1650) was the Other detected enzymes of this pathway without obvious peroxi- only conserved multi-pass membrane protein that had no clear assign- somal targeting signals were long-chain-fatty-acyl-CoA synthetase ment to another subcellular compartment, but the 3 identified peptides (EC 6.2.1.3) and thiolase (EC 2.3.1.16); these might have internal covered only 3% of the protein. Best matches to this sequence are signals or could be contaminants from another compartment. Mevalonate kinase is known to be in the glycosome41, and another enzyme of the mevalonate pathway, isopentenyl-diphosphate delta- Table 1. Known proteins of the glycosomal membrane. This list isomerase, was also present in our glycosomal preparation; but includes all mass-spectrometry-detected glycosomal membrane proteins. intermediate enzymes (phosphomevalonate kinase and mevalonate- 5-pyrophosphate decarboxylase) were not (Figure 2, Supplementary Lta gene ID Peptides Coverage Tb gene ID Function Table S1, sheet 2). Finally, we found a PTS1-containing phosphoribu- LtaP28.2340 16 59% Tb927.11.11520 PEX11 lokinase/uridine kinase family protein, LtaP14.0950, belonging to the LtaP24.0140 7 25% P-loop NTPase superfamily; we speculate that this enzyme might be involved in pyrimidine salvage. LtaP28.2330 3 14% LtaP25.2350 3 8% Tb927.3.2340 PEX2 3.4. Putative glycosomal membrane proteins LtaP15.0800 6 13% Tb927.9.6450 PEX16 We next focussed on the identification of putative membrane proteins. Known glycosomal membrane proteins and unassigned LtaP28.1050 3 15% Tb11.0300,Tb11.0400 proteins that had predicted membrane-spanning domains in both L. tarentolae and T. brucei are listed in Supplementary Table S1, LtaP26.2550 2 13% Tb927.9.1720 PMP4 sheet 3. These proteins were manually analysed for conserved LtaP35.3750 12 48% Tb927.9.11580, GIM5A,B functional domains, and their protein sequences were aligned to Tb927.9.11600 the complete predicted proteomes of Homo sapiens, Saccharomy- LtaP31.0560 5 5% Tb927.11.3130, GAT2 ces cerevisiae, Hansenula polymorpha and Pichia pastoris on the Tb927.4.4050 NCBI Web site using Blastp. This enabled the exclusion of further LtaP27.0470 4 6% Tb927.11.1070 GAT3 proteins that were deemed likely to be in other (non-peroxisomal) subcellular compartments. Abbreviations used: Lta, Leishmania tarentolae; Tb, Trypanosoma brucei. Page 7 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 Table 2. Possible additional glycosomal membrane proteins. This list includes all detected proteins with predicted trans-membrane (TM) domains, for which there is no evidence for location in any particular compartment. Lta gene ID Pep Cov Tb gene ID Function TMs TMs ER nr (%) (a) (b) sp LtaP29.1650 3 3 Tb927.3.4070 Major facilitator family, nitrate and chloride transporter? 12 nd no Tb927.3.4110 LtaP01.0560 4 17 Tb927.9.4310 putative tricarboxylate carrier, matches yeast putative alpha- 5 nd no isopropylmalate carrier, which exports alpha-isopropylmalate from the mitochondrion to the cytoplasm for use in leucine biosynthesis LtaP05.0670 2 6 Tb927.9.4310 putative tricarboxylate carrier, matches yeast putative alpha- 5 nd no isopropylmalate carrier, which exports alpha-isopropylmalate from the mitochondrion to the cytoplasm for use in leucine biosynthesis LtaP36.4130 3 6 Tb927.11.10260 hypothetical protein, conserved 0 2 no LtaP33.2680 4 4 Tb927.11.2590 hypothetical protein, conserved 0 2 no LtaP29.0320 2 21 Tb927.3.5350 hypothetical protein, conserved 1 1 no LtaP36.6980 4 4 Tb927.10.8000 P-loop nucleoside triphosphate hydrolase, contains EF hand domains 0 1 no LtaP21.0440 4 6 Tb927.10.2240 NTF2-like domain 0 1 no LtaP11.1210 2 5 Tb927.11.6070 ARM repeat superfamily 0 1 no LtaP31.3020 2 3 Tb927.4.5000, TERD-like domains 0 1 yes Tb927.8.7420 LtaP35.4380 2 4 Tb927.9.10470 hypothetical protein, conserved 0 1 no LtaP34.2250 2 5 Tb927.4.2250 hypothetical protein, conserved 0 1 no LtaP06.1080 3 11 Tb927.7.5700 hypothetical protein, conserved 1 nd yes LtaP26.0630 3 12 Tb927.7.1290 hypothetical protein, DUF2012, peptidase superfamily, starch binding 1 nd yes domain, similarity to human C15ORF24 LtaP36.3480 2 8 Tb927.11.10040 hypothetical protein, conserved 1 nd no LtaP17.1410 2 36 Tb927.5.2560 hypothetical protein, conserved 1 nd no LtaP23.0390 3 10 Tb927.8.2300 hypothetical protein, conserved 1 nd no Abbreviations used: Lta, Leishmania tarentolae; Tb, Trypanosoma brucei; Pep nr, number of identified peptides; Cov (%), percentage of protein coverage; TMs (a), number of annotated trans-membrane domains; TMs (b), number of transmembrane-domains identified by THMM TOP; ER sp, endoplasmic reticulum signal peptide. ion transporters. The Tb927.9.4310 protein sequence matches a yeast diacyl-phosphatidylcholine species, comprising common fatty ac- possible alpha-isopropylmalate carrier, which exports alpha-isopro- ids consisting of 16 to 22 carbon atoms with 0 to 6 desaturations. pylmalate from the mitochondrion to the cytoplasm for use in leucine In addition, membranes of intact trypanosomes also contained ether biosynthesis. The remaining candidates have either one or two poten- phospholipids species, both 1-alkyl-2-alkyl phosphatidylcholine tial trans-membrane domains, usually predicted by only one algorithm. species and 1-alkyl-2-acyl phosphatidylcholine species. The phos- phatidylcholine species composition of glycosomal membranes dif- Overall, our proteomics analysis revealed no clear candidates for fered only to a minor extent from the composition observed in total major novel glycosomal metabolite transporters. membranes of both bloodstream form and procyclic form trypano- somes (Table 3). 3.5. Phosphatidylcholine composition of cellular and glycosomal membranes 4. Discussion The phosphatidylcholine species composition in membranes from to- The glycosome is a major contributor to kinetoplastid energy metabo- tal T. brucei cells and glycosomes were analysed to detect possible dif- lism and essential for glycolysis3,4,47. Flux rates through the glycolytic ferences in membrane composition between glycosomal and the other pathway are high in trypanosomes and - judging from the peptide membranes of T. brucei, and to allow comparison in membrane com- counts and protein coverage in our analysis - the enzymes are also position between the two cultivatable replicating life cycle stages, the abundant in Leishmania. A model of trypanosome glycolysis that as- bloodstream form and procylic form. The phosphatidylcholine spe- sumes free exchange of glucose between the organelle and the cytosol cies composition of total trypanosome membranes differed to some mirrors the in vivo kinetics3,47. This suggests that the protein respon- extent between bloodstream and procyclic stages (Table 3), which sible for glucose transport should be very active and probably also is consistent with earlier reports on the phospholipid composition abundant. Our analysis of the glycosomal membrane proteome, how- in T. brucei46. Membranes of both stages contain predominantly ever, failed to identify abundant membrane proteins that might fulfil Page 8 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 Table 3. Phosphatidylcholine composition of procyclic-form and bloodstream-form T. brucei total cells and glycosomes. PCF BSF Peak Component Glycosomes Total cells Glycosomes Total cells Mol % Mol % Mol % Mol % 1 PtdCho 16:0, 16:1 2,9 1,7 ± 0,0 0,5 0,2 ± 0,2 2 AlkCho 16:0, 18:2 2,3 2,1 ± 0,3 0,7 nd 3 PtdCho 16:0, 18:3 3,3 1,7 ± 0,3 0,5 nd 4 PtdCho 16:0, 18:2 7,9 3,9 ± 0,5 2,5 3,9 ± 0,6 5 PtdCho 16:0, 18:1 2,8 3,1 ± 0,9 3,2 1,2 ± 0,3 6 AlkCho 16:0, 20:3 1,0 0,6 ± 0,0 0,2 nd 7 EnylCho 18:0, 18:2 0,9 1,0 ± 0,2 1,4 1,0 ± 0,2 8 AlkCho 18:0, 18:2 9,3 12,8 ± 0,7 9,6 6,6 ± 0,3 9 PtdCho 18:2, 18:2, 18:1, 18:3 12,6 6,1 ± 0,5 5,1 4,4 ± 0,4 10 PtdCho 18:1, 18:2 8,1 6,8 ± 0,2 4,3 5,6 ± 0,7 11 PtdCho 18:0, 18:2, 18:1, 18:1 23,2 13,4 ± 0,7 26,0 32,7± 4,6 12 PtdCho 18:0, 18:1 2,4 1,2 ± 0,2 2,4 2,2 ± 0,4 13 AlkCho 16:0, 22:1, 18:0, 20:1 0,5 0,6 ± 0,1 3,5 0,4 ± 0,7 14 PtdCho 16:0, 22:6 1,4 0,9 ± 0,1 nd nd 15 PtdCho 18:1, 20:5 2,8 3,9 ± 0,3 2,3 3,7 ± 0,3 16 PtdCho 16:0, 22:5 1,1 0,8 ± 0,0 1,1 2,1 ± 0,6 17 PtdCho 18:1, 20:4 1,3 2,6 ± 0,1 nd nd 18 PtdCho 16:0, 22:4 0,9 0,8 ± 0,2 2,3 2,2 ± 0,6 19 PtdCho 18:0, 20:3 2,2 1,8 ± 0,7 3,0 4,1 ± 0,4 20 PtdCho 18:3, 22:5 4,0 12,2 ± 0,1 4,9 4,5 ± 0,2 21 PtdCho 18:2, 22:5 0,3 nd 2,4 1,7 ± 0,3 22 PtdCho 18:1, 22:6 3,3 9,5 ± 2,1 3,5 3,5 ± 0,7 23 PtdCho 18:3, 22:3 0,4 0,9 ± 0,3 3,0 2,6 ± 1,0 24 PtdCho 18:0, 22:6 2,4 3,2 ± 0,3 3,2 5,3 ± 0,8 25 PtdCho 18:0, 22:5 2,6 5,5 ± 0,4 3,7 3,6 ± 0,3 26 PtdCho 18:2, 22:3 0,0 nd 3,2 3,4 ± 0,3 27 PtdCho 18:0, 22:4 0,3 0,5 ± 0,1 6,7 3,7 ± 0,4 28 PtdCho 18:0, 22:3 0,2 0,2 ± 0,1 0,9 1,3 ± 0,5 29 PtdCho 20:4, 22:6 0,0 1,2 ± 0,0 nd nd 30 PtdCho 20:4, 22:5 0,0 1,0 ± 0,1 nd nd The phosphatidylcholine species description comprises the sn-1 linkage type followed by the radyl chains on the sn-1 and sn-2 position, respectively. “Total cell” values are mean of three independent experiments. Most abundant species representing over 5 Mol % are marked in bold. Abbreviations: AlkCho, 1-alkyl, 2-acyl phosphatidylcholine; BSF, bloodstream-form T. brucei; EnylCho, 1-alkyl-1-enyl-2-acyl phosphatidylcholine; nd, not detected; PCF, procyclic-form T. brucei; PtdCho, diacyl phosphatidylcholine. such a role. Given the large number of proteins that we identified cannot be ruled out. If so, they could possibly function as putative gly- - including multiple membrane proteins from other compartments - it cosomal isocitrate/2-ketoglutarate and fumarate/malate shuttles. seems unlikely that our failure to detect transporters for such major metabolites could be due solely to lack of sensitivity. Dual subcellular We investigated the lipid composition of glycosomal membranes in locations could be a possible explanation for some proteins, as pre- T. brucei by analysis of the species composition of phosphatidylcho- viously shown for MCP617. The five additional mitochondrial carrier line, the most abundant phospholipid class in membranes of both pro- proteins that we detected are predominantly in the mitochondrion of cyclic and bloodstream form T. brucei46. The phospholipid composition procyclic T. brucei48, but the presence of a minor amount in glycosomes of peroxisomal membranes has been investigated in peroxisomes Page 9 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 isolated from rat liver and from the yeasts Saccharomyces cerevisiae to exist in kinetoplastids. Instead, it has been speculated that PEX11 and Pichia pastoris49–51. In rat liver peroxisomes the phospholipid family proteins might contribute to glycosomal porin activity5,8. Our classes and their fatty acid composition were similar to those of ho- failure to find abundant novel glycosomal transporters is consistent mogenates and microsomes, although exposure to the endotoxin li- with the hypothesis that the PEX11 family proteins are indeed re- popolysaccharide (LPS) induced significant changes in phospholipid sponsible for the transfer of small solutes in and out of the glycosome. species composition: in particular, the abundance of both long chain fatty acids (>20 C atoms) and poly-unsaturated fatty acids increased in peroxisomal membranes49. The lipid composition of peroxisomes in yeasts was shown to be rather flexible, predominantly depending Author contributions on the type and amount of fatty acid supply in the medium50,51. The C. Colasante was responsible for the glycosome and glycosomal lipid composition of glycosomes in Trypanosomatidae has not been membrane purification, with supervision from FV and C. Clayton. investigated before and our results showed that the phosphatidylcho- C. Colasante was responsible for biochemical pathway analysis. line species composition in glycosomal membranes resembled that JvH and AGMT were responsible for the lipid analysis, and TR for of other cellular membranes in both bloodstream-form and procy- the mass spectrometry. C. Colasante, TM and C. Clayton contrib- clic-form T. brucei. These results suggest that the lipid composition, uted bioinformatic analysis of the identified proteins. The paper and thus the biophysical properties, of the glycosomal membrane is was written by C. Colasante, FV, JvH and C. Clayton. similar to that of the other membranes in trypanosomes. Because of this similarity to other cellular membranes, glycosomal membranes Competing interests are expected to be also impermeable to small hydrophilic molecules The authors have no competing interests to declare. unless transport is facilitated by specialized membrane proteins. Grant information There is accumulating evidence that small solutes enter microbodies Work by C. Colasante and F. Voncken was supported by the through pores5,20,52,53. Evidence from a mammalian Pxmp2 (PMP22) Deutsche Forschungsgemeinschaft (DFG) (Cl112/10). knock-out mouse suggested loss of peroxisomal pores for solutes of under 300 Da53 and this type of function was confirmed when the The funders had no role in study design, data collection and analy- protein was expressed in insect cells. No Pxmp2 homologue appears sis, decision to publish, or preparation of the manuscript. Page 10 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 Supplementary figure Myc Mitotracker Merge - + tet kDa Tb927.10.14020 48 Tb927.3.1840 35 Tb927.5.1210 34 Tb927.11.1270 20 Tb927.10.520 40 Myc Bip Merge - + tet kDa Tb927.7.3900 20 Supplementary Figure S1. Immunofluorescence analysis of the putative glycosomal membrane proteins in procyclic T. brucei. N-terminally or C-terminally myc-tagged versions of the proteins were detected (green) by using a commercial Myc antibody (Sigma-Aldrich). Mitochondria were visualized (red) using mitotracker. The endoplasmatic reticulum (ER) was detected using an antibody directed against the ER lumen protein BiP (red). Overlays (Merge) of the green staining and the red staining are shown to visualize the common compartmentalization of the proteins. On the right side, western blots are shown to illustrate the expression of the myc-tagged membrane proteins. (+) and (-) indicate tetracycline-induced and -uninduced cells respectively. Page 11 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 References 1. Colasante C, Ellis M, Ruppert T, et al.: Comparative proteomics of glycosomes 21. Alibu VP, Storm L, Haile S, et al.: A doubly inducible system for RNA from bloodstream form and procyclic culture form Trypanosoma brucei interference and rapid RNAi plasmid construction in Trypanosoma brucei. brucei. Proteomics. 2006; 6(11): 3275–3293. Mol Biochem Parasitol. 2005; 139(1): 75–82. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 2. Vertommen D, Van Roy J, Szikora JP, et al.: Differential expression of glycosomal 22. Hirumi H, Hirumi K: Continuous cultivation of Trypanosoma brucei bloodstream and mitochondrial proteins in the two major life-cycle stages of Trypanosoma forms in a medium containing a low concentration of serum protein without brucei. Mol Biochem Parasitol. 2008; 158(2): 189–201. feeder cell layers. J Parasitol. 1989; 75(6): 985–989. PubMed Abstract | Publisher Full Text PubMed Abstract 3. Haanstra JR, van Tuijl A, van Dam J, et al.: Proliferating bloodstream-form 23. van Deursen FJ, Shahi SK, Turner CM, et al.: Characterisation of the growth and Trypanosoma brucei use a negligible part of consumed glucose for anabolic differentiation in vivo and in vitro of bloodstream-form Trypanosoma brucei processes. Int J Parasitol. 2012; 42(7): 667–673. strain TREU 927. Mol Biochem Parasitol. 2001; 112(2): 163–171. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 4. Bakker BM, Mensonides FI, Teusink B, et al.: Compartmentation protects 24. Raymond F, Boisvert S, Roy G, et al.: Genome sequencing of the lizard parasite trypanosomes from the dangerous design of glycolysis. Proc Natl Acad Sci Leishmania tarentolae reveals loss of genes associated to the intracellular U S A. 2000; 97(5): 2087–2092. stage of human pathogenic species. Nucleic Acids Res. 2012; 40(3): 1131–1147. PubMed Abstract | Publisher Full Text | Free Full Text PubMed Abstract | Publisher Full Text | Free Full Text 5. Gualdron-López M, Vapola MH, Miinalainen IJ, et al.: Channel-forming 25. Neuberger G, Maurer-Stroh S, Eisenhaber B, et al.: Prediction of peroxisomal activities in the glycosomal fraction from the bloodstream form of targeting signal 1 containing proteins from amino acid sequence. J Mol Biol. Trypanosoma brucei . PLoS One. 2012; 7(4): e34530. 2003; 328(3): 581–592. PubMed Abstract | Publisher Full Text | Free Full Text PubMed Abstract | Publisher Full Text 6. Aman RA, Wang CC: Identification of two integral glycosomal membrane 26. Neuberger G, Maurer-Stroh S, Eisenhaber B, et al.: Motif refinement of the proteins in Trypanosoma brucei. Mol Biochem Parasitol. 1987; 25(1): 83–92. peroxisomal targeting signal 1 and evaluation of taxon-specific differences. PubMed Abstract | Publisher Full Text J Mol Biol. 2003; 328(3): 567–579. 7. Lorenz P, Maier AG, Erdmann R, et al.: Elongation and clustering of glycosomes PubMed Abstract | Publisher Full Text in Trypanosoma brucei overexpressing the glycosomal Pex11p. EMBO J. 1998; 27. Hirokawa T, Boon-Chieng S, Mitaku S: SOSUI: classification and secondary 17(13): 3542–3555. structure prediction system for membrane proteins. Bioinformatics. 1998; PubMed Abstract | Publisher Full Text | Free Full Text 14(4): 378–379. 8. Voncken F, van Hellemond JJ, Pfisterer I, et al.: Depletion of GIM5 causes cellular PubMed Abstract | Publisher Full Text fragility, a decreased glycosome number, and reduced levels of ether-linked 28. Tusnády GE, Simon I: The HMMTOP transmembrane topology prediction phospholipids in trypanosomes. J Biol Chem. 2003; 278(37): 35299–35310. server. Bioinformatics. 2001; 17(9): 849–850. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 9. Maier A, Lorenz P, Voncken F, et al.: An essential dimeric membrane protein of 29. Bligh EG, Dyer WJ: A rapid method of total lipid extraction and purification. trypanosome glycosomes. Mol Microbiol. 2001; 39(6): 1443–1451. Can J Biochem Physiol. 1959; 37(8): 911–917. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 10. Vertommen D, Van Roy J, Szikora JP, et al.: Differential expression of 30. Ferro M, Seigneurin-Berny D, Rolland N, et al.: Organic solvent extraction as a glycosomal and mitochondrial proteins in the two major life-cycle stages of versatile procedure to identify hydrophobic chloroplast membrane proteins. Trypanosoma brucei. Mol Biochem Parasitol. 2008; 158(2): 189–201. Electrophoresis. 2000; 21(16): 3517–3526. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 11. Flaspohler JA, Rickoll WL, Beverley SM, et al.: Functional identification 31. Da Cruz S, Xenarios I, Langridge J, et al.: Proteomic analysis of the mouse liver of a Leishmania gene related to peroxin 2 reveals common ancestry of mitochondrial inner membrane. J Biol Chem. 2003; 278(42): 41566–41571. glycosomes and peroxisomes. Mol Cell Biol. 1997; 17(3): 1093–1101. PubMed Abstract | Publisher Full Text PubMed Abstract | Free Full Text 32. Quinones W, Urbina JA, Dubourdieu M, et al.: The glycosome membrane 12. Guerra-Giraldez C, Quijada L, Clayton CE: Compartmentation of enzymes in of Trypanosoma cruzi epimastigotes: protein and lipid composition. a microbody, the glycosome, is essential in Trypanosoma brucei. J Cell Sci. Exp Parasitol. 2004; 106(3–4): 135–149. 2002; 115(Pt 13): 2651–8. PubMed Abstract | Publisher Full Text PubMed Abstract 33. Soualhine H, Brochu V, Menard F, et al.: A proteomic analysis of 13. Krazy H, Michels PA: Identification and characterization of three peroxins– penicillin resistance in Streptococcus pneumoniae reveals a novel PEX6, PEX10 and PEX12–involved in glycosome biogenesis in Trypanosoma role for PstS, a subunit of the phosphate ABC transporter. Mol Microbiol. brucei. Biochim Biophys Acta. 2006; 1763(1): 6–17. 2005; 58(5): 1430–1440. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 14. Brennand A, Rigden DJ, Michels PA: Trypanosomes contain two highly different 34. Panigrahi AK, Ogata Y, Zíková A, et al.: A comprehensive analysis of isoforms of peroxin PEX13 involved in glycosome biogenesis. FEBS Lett. Trypanosoma brucei mitochondrial proteome. Proteomics. 2008; 9(2): 434–450. 2012; 586(13): 1765–71. PubMed Abstract | Publisher Full Text | Free Full Text PubMed Abstract | Publisher Full Text 35. Opperdoes FR, Szikora JP: In silico prediction of the glycosomal enzymes of 15. Moyersoen J, Choe J, Kumar A, et al.: Characterization of Trypanosoma brucei Leishmania major and trypanosomes. Mol Biochem Parasitol. 2006; 147(2): PEX14 and its role in the import of glycosomal matrix proteins. Eur J Biochem. 193–206. 2003; 270(9): 2059–67. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text 36. Bandini G, Marino K, Guther ML, et al.: Phosphoglucomutase is absent in 16. Yernaux C, Fransen M, Brees C, et al.: Trypanosoma brucei glycosomal ABC Trypanosoma brucei and redundantly substituted by phosphomannomutase transporters: identification and membrane targeting. Mol Membr Biol. 2006; and phospho-N-acetylglucosamine mutase. Mol Microbiol. 2012; 85(3): 513–534. 23(2): 157–72. PubMed Abstract | Publisher Full Text | Free Full Text PubMed Abstract | Publisher Full Text 37. Galland N, de Walque S, Voncken FG, et al.: An internal sequence targets 17. Colasante C, Alibu VP, Kirchberger S, et al.: Characterization and Trypanosoma brucei triosephosphate isomerase to glycosomes. Mol Biochem developmentally regulated localization of the mitochondrial carrier Parasitol. 2010; 171(1): 45–49. protein homologue MCP6 from Trypanosoma brucei. Eukaryotic Cell. 2006; PubMed Abstract | Publisher Full Text 5(8): 1194–1205. 38. Penha LL, Sant'Anna CB, Mendonca-Previato L, et al.: Sorting of PubMed Abstract | Publisher Full Text | Free Full Text phosphoglucomutase to glycosomes in Trypanosoma cruzi is 18. Islinger M, Luers GH, Zischka H, et al.: Insights into the membrane proteome of mediated by an internal domain. Glycobiology. 2009; 19(12): 1462–1472. rat liver peroxisomes: microsomal glutathione-S-transferase is shared by both PubMed Abstract | Publisher Full Text subcellular compartments. Proteomics. 2006; 6(3): 804–816. 39. Zomer AW, Michels PA, Opperdoes FR: Molecular characterisation of PubMed Abstract | Publisher Full Text Trypanosoma brucei alkyl dihydroxyacetone-phosphate synthase. 19. Wiese S, Gronemeyer T, Ofman R, et al.: Proteomics characterization of mouse Mol Biochem Parasitol. 1999; 104(1): 55–66. kidney peroxisomes by tandem mass spectrometry and protein correlation PubMed Abstract | Publisher Full Text profiling. Mol Cell Proteomics. 2007; 6(12): 2045–2057. 40. Wiemer EA, IJlst L, van Roy J, et al.: Identification of 2-enoyl coenzyme PubMed Abstract | Publisher Full Text A hydratase and NADP(+)-dependent 3-hydroxyacyl-CoA dehydrogenase 20. Visser WF, van Roermund CW, Ijlst L, et al.: Metabolite transport across the activity in glycosomes of procyclic Trypanosoma brucei. Mol Biochem peroxisomal membrane. Biochem J. 2007; 401(12): 365–375. Parasitol. 1996; 82(1): 107–111. PubMed Abstract | Publisher Full Text | Free Full Text PubMed Abstract | Publisher Full Text Page 12 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 41. Carrero-Lerida J, Perez-Moreno G, Castillo-Acosta VM, et al.: Intracellular Trypanosoma brucei can be understood in terms of the kinetics of the location of the early steps of the isoprenoid biosynthetic pathway in glycolytic enzymes. J Biol Chem. 1997; 272(6): 3207–3215. the trypanosomatids Leishmania major and Trypanosoma brucei. Int J PubMed Abstract | Publisher Full Text Parasitol. 2009; 39(3): 307–314. 48. Colasante C, Peña Diaz P, Clayton C, et al.: Mitochondrial carrier PubMed Abstract | Publisher Full Text family inventory of Trypanosoma brucei brucei: Identification, 42. Matsuzaki T, Fujiki Y: The peroxisomal membrane protein import receptor expression and subcellular localisation. Mol Biochem Parasitol. Pex3p is directly transported to peroxisomes by a novel Pex19p- and 2009; 167(2): 104–17. Pex16p-dependent pathway. J Cell Biol. 2008; 183(7): 1275–1286. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text | Free Full Text 49. Khan M, Contreras M, Singh I: Endotoxin-induced alterations of lipid and fatty 43. Alsford S, Turner DJ, Obado SO, et al.: High throughput phenotyping using acid compositions in rat liver peroxisomes. J Endotoxin Res. 2000; 6(1): 41–50. parallel sequencing of RNA interference targets in the African trypanosome. PubMed Abstract | Publisher Full Text Genome Res. 2011; 21(6): 915–924. 50. Tuller G, Nemec T, Hrastnik C, et al.: Lipid composition of subcellular PubMed Abstract | Publisher Full Text | Free Full Text membranes of an FY1679-derived haploid yeast wild-type strain grown on 44. Reguenga C, Oliveira ME, Gouveia AE, et al.: Identification of a 24 kDa intrinsic different carbon sources. Yeast. 1999; 15(14): 1555–1564. membrane protein from mammalian peroxisomes. Biochim Biophys Acta. 1999; PubMed Abstract | Publisher Full Text 1445(3): 337–341. 51. Wriessnegger T, Gübitz G, Leitner E, et al.: Lipid composition of peroxisomes PubMed Abstract | Publisher Full Text from the yeast Pichia pastoris grown on different carbon sources. Biochim 45. Marshall PA, Krimkevich YI, Lark RH, et al.: Pmp27 promotes peroxisomal Biophys Acta. 2007; 1771(4): 455–461. proliferation. J Cell Biol. 1995; 129(129): 345–355. PubMed Abstract | Publisher Full Text PubMed Abstract | Publisher Full Text | Free Full Text 52. Antonenkov VD, Hiltunen JK: Peroxisomal membrane permeability and solute 46. Patnaik PK, Field MC, Menon AK, et al.: Molecular species analysis of transfer. Biochim Biophys Acta. 2006; 1763(12): 1697–1706. phospholipids from Trypanosoma brucei bloodstream and procyclic forms. PubMed Abstract | Publisher Full Text Mol Biochem Parasitol. 1993; 58(1): 97–105. 53. Rokka A, Antonenkov VD, Soininen R, et al.: Pxmp2 is a channel-forming protein PubMed Abstract | Publisher Full Text in mammalian peroxisomal membrane. PLoS One. 2009; 4(4): e5090. 47. Bakker BM, Michels PA, Opperdoes FR, et al.: Glycolysis in bloodstream-form PubMed Abstract | Publisher Full Text | Free Full Text Page 13 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 Open Peer Review Current Referee Status: Version 1 Referee Report 27 March 2013 doi:10.5256/f1000research.1189.r868 Peter J. Myler Seattle Biomedical Research Institute, Seattle, WA, USA In this manuscript, Colasante et al. describe an exhaustive analysis of the membrane proteins and lipids present in glycosomes from the kinetoplastid protozoan parasite, Leishmania tarentolae, and conclude that, in the absence of obvious transporters for the major substrates and intermediates of energy metabolism, these molecules must be transported across the glycosomal membrane via pores formed by known glycosomal membrane proteins, which are orthologues of the peroxisomal membrane proteins PEX11, GIM5A/B, PXMP4, PEX2 and PEX16. The paper is well-written and the technical approach is appropriate. While one must be mindful of the adage “absence of evidence is not evidence of absence”, the comprehensive nature of the proteomic analysis in the present study provides a persuasive argument (albeit, not definitive proof) for this conclusion. As a minor point, it could perhaps have been helpful to have indicated the identity of the proteins enriched in the glycosomal membrane fraction (urea pellet) of Table 1B. It would also be good to indicate what proportions of the total glycosomal fraction are represented by the two lanes in this figure. I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Competing Interests: No competing interests were disclosed. Referee Report 08 February 2013 doi:10.5256/f1000research.1189.r758 Ralf Erdmann Department of Biochemistry Systems, Institute of Physiological Chemistry, Ruhr-Universität Bochum, Bochum, Germany Glycosomes are evolutionarily and functionally related to peroxisomes. To fulfill their metabolic functions, the organelles communicate with the cytosol by an exchange of metabolites/products. Driven by the fact that the knowledge on metabolite transport across the glycosomal membrane and the nature of involved protein in this process is still scarce, the authors performed a proteomic approach with purified organelles. The paper is well written, the experimental design and results are conclusive. F1000Research Page 14 of 15 F1000Research 2013, 2:27 Last updated: 25 DEC 2016 The paper is well written, the experimental design and results are conclusive. Some minor comments should be addressed: - Figure 1A: Please indicate top/bottom fractions and provide some information of the density at least of the peak fractions. Please mention in the legend how much of the total fractions was subjected to the gel. - Figure 1B: Please indicate the load. Are the membranes enriched compared to the glycosomal fraction? Please also provide more details on the experimental approach in the methods section, like volumes used for resuspension, fractions loaded on the gel, etc. I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Competing Interests: No competing interests were disclosed. F1000Research Page 15 of 15