Rab23 is a flagellar protein in Trypanosoma brucei
© Lumb et al; licensee BioMed Central Ltd. 2011
Received: 1 December 2010
Accepted: 15 June 2011
Published: 15 June 2011
Skip to main content
We’re sorry, something doesn't seem to be working properly. Please try refreshing the page. If that doesn't work, please contact us so we can address the problem.
© Lumb et al; licensee BioMed Central Ltd. 2011
Received: 1 December 2010
Accepted: 15 June 2011
Published: 15 June 2011
Rab small GTPases are important mediators of membrane transport, and orthologues frequently retain similar locations and functions, even between highly divergent taxa. In metazoan organisms Rab23 is an important negative regulator of Sonic hedgehog signaling and is crucial for correct development and differentiation of cellular lineages by virtue of an involvement in ciliary recycling. Previously, we reported that Trypanosoma brucei Rab23 localized to the nuclear envelope , which is clearly inconsistent with the mammalian location and function. As T. brucei is unicellular the potential that Rab23 has no role in cell signaling was possible. Here we sought to further investigate the role(s) of Rab23 in T. brucei to determine if Rab23 was an example of a Rab protein with divergent function in distinct taxa.
The taxonomic distribution of Rab23 was examined and compared with the presence of flagella/cilia in representative taxa. Despite evidence for considerable secondary loss, we found a clear correlation between a conventional flagellar structure and the presence of a Rab23 orthologue in the genome. By epitope-tagging, Rab23 was localized and found to be present at the flagellum throughout the cell cycle. However, RNAi knockdown did not result in a flagellar defect, suggesting that Rab23 is not required for construction or maintenance of the flagellum.
The location of Rab23 at the flagellum is conserved between mammals and trypanosomes and the Rab23 gene is restricted to flagellated organisms. These data may suggest the presence of a Rab23-mediated signaling mechanism in trypanosomes.
Rab proteins are important control elements of vesicle transport and related functions . In African trypanosomes there are sixteen Rab-like proteins with various roles in exocytosis, endocytosis and possibly differentiation/life cycle progression [3–5]. Essentially members of the family have similar locations and functions to their mammalian or fungal orthologues where such orthologous relationships exist [4, 6].
In contrast, amongst the Trypanosoma brucei Rab repertoire is Rab23 (TbRab23), which was localized to the inner nuclear envelope using polyclonal antibodies . Significantly, mammalian Rab23 was first recognized in a murine model through genetic interactions with Sonic hedgehog (Shh), a protein crucial to development and differentiation of cell lineages [7, 8]. Mammalian Rab23 localizes to the plasma membrane, endosomes, and cytosol and functions as a negative regulator of Shh [9–11].
The Shh pathway is metazoan-specific and essential for growth and organ patterning [12, 13]. Several components of the pathway localize to primary cilia and intraflagellar transport (IFT) is required for Shh signaling [14–18]. Recent data indicate that Rab23 is essential for ciliogenesis, localizes to cilia in MDCK cells and functions to recycle the Shh pathway receptor, Smoothened (Smo), from the ciliary compartment [19, 20].
As Shh signaling is restricted to metazoa, this raises intriguing questions about the possible function of TbRab23 in T. brucei and other non-metazoan organisms and may point towards a divergent function. Given that no Rab protein has been associated with the nucleus in any organism since our earlier publication, and accumulation of evidence for a role of Rab23 at the mammalian flagellum, we sought to re-examine the location of Rab23 in trypanosomes and its phylogenetic relationship with the presence of a flagellum. Using epitope-tagged chimeras of TbRab23 to avoid issues with antisera cross-reactivity, we now find that TbRab23 is associated with the flagellum, and not the nucleus as previously reported . Further, Rab23 is only found in flagellated taxa. Importantly our data are consistent with the studies in mammalian cells and indicates that the function of Rab23 is likely conserved across the eukaryotes.
Genomes were chosen for sampling based on three main criteria: i) proposed position in the evolutionary tree of eukaryotes  ensuring adequate representation of major supergroups, ii) completeness of the sequenced genome, and iii) the availability of a carefully annotated protein database. BLAST searches were conducted initially using TbRab23 (Tb10.6k15.1990), TbRab28 (Tb927.6.3040), TbRab7 (Tb09.211.2330), HsRab23 (NP_057361.3), and HsRab28 (CAA64364.1) as queries . As GTPases are very highly conserved, e values of less than e-20 were used as an initial cutoff. The top two hits were then reciprocally used as queries for a BLAST search against T. brucei and Homo sapiens sequence databases. Only those that retrieved the query sequence were assigned as candidate orthologues. If no clear orthologue was identified, additional searches were performed using protein sequences from closely related species. In cases where the retrieved sequence was a named orthologue of the query, reciprocal BLAST searches were not performed. The Rab5 dataset was kindly provided by Dr. Joel Dacks, University of Calgary.
Initial protein sequence alignments were produced in Clustal ×  on the ODIN server at the University of Calgary, and manually edited in MacClade . Phylogenetic reconstructions were performed by analyzing the alignment using ProtTest (V1.3)  to select the appropriate model of sequence evolution, and then MrBayes (V3.2) run locally for 106 generations and PhyML at http://www.atgc-montpellier.fr/phyml (100 pseudoreplicates) under WAG+I+G substitution model .
Trypanosoma brucei brucei bloodstream form (BSF) and procyclic form (PCF) Lister 427 strains were maintained in HMI-9 and SDM79 medium respectively supplemented with 10% foetal calf serum (FCS), L-glutamine and penicillin/streptomycin as described previously . Single Marker BSF (SMB) cells  were cultured in HMI-9 with the addition of geneticin to a final concentration of 2.5 μg/ml. Tetracycline-responsive PTT PCF  were cultured with the addition of hygromycin and geneticin, each to a final concentration of 25 μg/ml. Cells in exponential growth phase ( < 1 × 106/ml and 3-8 × 106 cells/ml for BSF and PCF cells respectively) were used in all experiments.
Primers were designed and verified for specificity using RNAit  to amplify a fragment of the TbRab23 ORF from BSF and PCF genomic DNA (BSF: For- CGCTCTCGAGGAGGTGGGTTAAACGGAGTG, rev- GCTAGGATCCCGCTCGGTCCACATTC; PCF: For- GCATGGATCCAAGAGCGTTTGCCTCCGTCA, rev- GCTTAAGCTTCCGCTCCACCAGATTTCGAT). The fragments were cloned into the tetracycline-inducible p2T7TABlue or the p2T7177 plasmid for RNAi in BSF and PCF respectively. An AMAXA Nucleofector® II was used following the manufacturer's protocol as described previously  to transfect tetracycline-responsive SMB cells with NotI-digested plasmid. Single clones were selected and maintained in the presence of 5 μg/ml hygromycin and 2.5 μg/ml geneticin. PTT PCF stable cell lines were generated in cytomix (pH 7.6 2 mM EGTA, 120 mM KCl, 0.15 mM CaCl2, pH 7.6 10 mM K2HPO4/KH2PO4, pH 7.6 25 mM HEPES, 5 mM MgCl2.6H2O, 0.5% Dextrose, 100 μg/ml BSA, 1 mM Hypoxanthine) containing 20 μg linearized DNA using a Bio-Rad Gene Pulser II set to 1.5 kv and a capacitance of 25 μF. Single clones were selected by serial dilution in medium containing 20% FCS, 25 μg/ml hygromycin, 25 μg/ml geneticin, and 3 μg/ml phleomycin.
For ectopic expression, TbRab23 was amplified from genomic DNA (For: CGGAAGCTTCGTTGAAGAGAGGTGGGTTAAAC, rev: GCTACGGGATCCCTACATAACACTGCACTTTTTCTTC). The PCR product was cloned into either pHD1034 or pXS219 containing an N-terminal HA- or YFP-epitope, for expression in BSF and PCF cells respectively. WT SMB cells or SMB cells containing the pT7TABlue-TbRab23 plasmid were transfected with 5 μg of linearized pHD1034-TbRab23. SMB pHD1034-TbRab23 single clones were selected and maintained in the presence of 2.5 μg/ml geneticin and 0.2 μg/ml puromycin, SMB cells transgenic for both TbRab23 RNAi and expression were selected and maintained as above with the addition of 5 μg/ml hygromycin. Twenty micrograms of linearized pXS219-TbRab23 was used to transfect WT PTT cells or PTT p2T7177-TbRab23 cells. PTT pXS219-TbRab23 single clones were selected in medium containing 25 μg/ml hygromycin, 25 μg/ml geneticin, and 1.5 μg/ml puromycin. PTT cells transgenic for both TbRab23 RNAi and expression were selected and maintained as above with the addition of 3 μg/ml phleomycin.
RNAi was induced in log-phase parasites by the addition of tetracycline at 1 μg/ml. Silencing of the target was monitored by Western blotting on cell lines expressing epitope-tagged TbRab23. BSF proliferation curves were performed in triplicate by inoculating cultures at 1 × 104 cells/ml each day in fresh medium with antibiotics. Cell densities were determined using a Coulter Z1 Counter (Beckman). PCF proliferation curves were performed in triplicate by inoculating cultures at 5 × 105 cells/ml every two days in fresh medium with antibiotics.
Cells were harvested at 800 g for 10 minutes and washed twice in ice-cold phosphate buffered saline (PBS) (Sigma). 107 cells were heated in 2× SDS-PAGE loading buffer (100 mM Tris-HCl, pH 6.8, 20% v/v glycerol, 0.2% v/v β-mercaptoethanol, 0.2% w/v bromophenol blue, 4% w/v SDS) for 5 minutes at 94°C. 107 cells per lane were loaded and resolved by 12.5% SDS-PAGE. Proteins were electrophoretically transferred onto polyvinylidene fluoride membranes (Millipore). Membranes were blocked and processed following standard procedures. Polyclonal anti-TbBiP serum (a kind gift from J. D. Bangs, University of Wisconsin) was used at 1:20000, polyclonal anti-GFP (a gift from M. Rout, Rockefeller University) 1:10000, polyclonal anti-TbRab11 1:1000, monoclonal anti-PFR (L8C4, a gift from K. Gull, University of Oxford) 1:1000, monoclonal anti-β tubulin (Chemicon) 1:10000, TbRab23 peptide sera 1:1000, and monoclonal anti-HA (Roche) 1:5000. Incubations with the appropriate commercial secondary anti-IgG horseradish peroxidase (HRP) conjugates (Sigma) were performed at 1:20000 in 1x Tris-buffered saline (with 0.1% tween 20) for 45 minutes. Detection was by chemiluminescence with luminol (Sigma) on BioMaxMR film (Kodak).
Flagella isolation was performed as previously described . Briefly, 1 × 108 cells were lysed in PMN buffer (10 mM NaPO4, pH 7.2, 150 mM NaCl, 1 mM MgCl2, mini-complete, EDTA-free, protein inhibitor tablets (Roche), 0.5% Triton X-100 (v/v)) and subpellicular microtubules depolymerized by treatment with PMN buffer supplemented with 1 M sodium chloride. Flagella were isolated by centrifugation at 16000 g for 10 minutes. Supernatant fractions were concentrated by TCA precipitation using standard procedures and each fraction was subjected to SDS-PAGE and immunoblotting. Whole cell equivalents were loaded. All centrifugation and incubation steps were carried out at 4°C.
1× 108 BSF and 5 × 107 PCF cells were harvested at 800 g for 10 minutes at 4°C and total RNA extracted using the RNeasy mini kit (Qiagen) according to the manufacturer's instructions. cDNA was generated from 2 μg RNA using a SuperscriptTM II RNaseH reverse transcriptase kit (Invitrogen). Quantitative RT-PCR was performed using iQ-SYBRGreen Supermix on a Mini Opticon Real-time PCR Detection System (BioRad) and quantified using Opticon3 software (BioRad). β-tubulin is expressed at similar levels between BSF and PCF cells  and used to normalize RNA input. A melting curve and gel electrophoresis was performed on amplified products to ensure amplification of the desired fragment. The following primers were used: TbRab23 F-AGTGGCTGAAGAGGTGGAGA, TbRab23 R-CACATTGCTGCCACGTACTC, β-tubulin F-CAAGATGGCTGTCACCTTCA and β-tubulin R-GCCAGTGTACCAGTGCAAGA.
Cells were grown to log phase, fixed in 3% (w/v) paraformaldehyde in PBS on ice and adhered to poly-L-lysine slides. Cells were permeabilized by incubating in 0.1% (v/v) Triton-X-100, washed and blocked in 20% v/v foetal calf serum. Slides were incubated with primary antibodies for 1 hour and washed. Secondary antibodies were applied at 1:1000 for 1 hour at ambient temperature. Slides were mounted with Vectashield containing 4',6-diamidino-2-phenylindole (DAPI) to stain DNA (Vectalabs) and visualized on a Nikon Eclipse E600 epifluorescence microscope fitted with a Hamamatsu CCD digital camera. Image acquisition was performed with Metamorph software (Molecular Devices) and processing with Photoshop (Adobe). All images were taken under non-saturating conditions. All quantitation was done using identical exposures as appropriate and using the raw data within Metamorph. Subsequent image processing was performed for presentation purposes with Photoshop; representative images are shown. Antibodies were used at the following dilutions: TbPFR (L8C4) 1:50, FAZ (L3B2, a gift from K. Gull) 1:5, anti-HA 1:2000, and anti-GFP 1:2500. Isolated flagella were prepared for immunofluorescence analysis in solution, omitting the permeabilization step, and were spotted onto slides after the staining protocol was complete.
1x108 BSF parasites expressing TbRab23YFP were harvested by centrifugation for 10 minutes at 800 g and washed with saline (0.1 M HEPES pH 7.0, 0.98% sodium chloride). Cells were fixed in 2% formaldehyde and 0.05% glutaraldehyde in 0.1 M PIPES for 1 hour. After 3 washes in 0.1 M HEPES pH 7.0, the cell pellet was infiltrated overnight in 2.3 M sucrose (in PBS) at 4°C. Specimens were frozen in liquid nitrogen and 65 nm ultrathin cryosections cut on a Leica UCT ultramicrotome with EM FCS cryoattachment. Sections were mounted on grids with formvar support, blocked with 2 mM glycine (in PBS) followed by 10%FCS (in PBS). Anti-GFP antibody was applied for 1 hour, and the samples washed three times in PBS (5 minutes each). 15 nm Protein A gold was then applied for 30 minutes followed by three rinses in PBS. Six further washes were performed in dH2O over a 5-minute period and specimens were prepared for visualization by applying uranyl acetate in methylcellulose on ice for 10 minutes. Sections were viewed using a Philips CM100 electron microscope (FEI-Philips) operated at 80 Kv.
All materials used in this study were obtained from accredited sources who operate ethical business practices.
As mammalian Rab23 has some function at the endosome , the relationship between Rab23 and other major endocytic Rab proteins, Rab5, Rab7 and Rab28 (Lumb et al., 2011, submitted) was investigated. Phylogenetic reconstruction of Rab23 unequivocally demonstrated that Rab23 is monophyletic with high statistical support and consistent with previous reconstructions (Figure 1B) [1, 10, 38]. Furthermore, correlation between the presence of a Rab23 gene and possession of a motile flagellum was found (Figure 1C). Out of 26 species sampled, 15 have motile flagella, and, bar Giardia lamblia and Plasmodium falciparum both of which have non-canonical flagella, all possess a Rab23 homologue. Rab23 orthologues were not found in any of the non-flagelated organisms, suggesting that Rab23 has flagellum-associated functions, supported by the finding that dominant negative forms of Rab23 prevent primary cilia formation in mammalian cells .
Given the increasing evidence for flagellar-associated functions for mammalian Rab23 and the evolutionary correlation between Rab23 and motile flagella, we sought to re-examine the location of TbRab23. Re-investigation of both the original antisera and a new anti-peptide sera suggested significant issues with specificity (data not shown and Additional file 2). For example, we observed high molecular weight cross-reacting bands in blots using the original antisera, which may represent detection of highly repetitive polypeptides at the nuclear envelope , and the peptide sera did not recognise endogenous TbRab23 by Western blot or IF. Given the low level of expression of TbRab23 as a confounding factor and questionable specificity of antisera, we chose an independent strategy to re-examine TbRab23 location.
Production of ectopic TbRab23HA (27 kDa) and TbRab23YFP (51 kDa) proteins of the correct molecular weight were verified by Western blotting (Figure 2B). The location of epitope-tagged TbRab23 was followed throughout the trypanosome cell cycle. Using an anti-GFP antibody, TbRab23YFP was localized to the cell body, an elongated structure and a punctate feature adjacent to the kinetoplast (Figure 2D, left panel). The positioning and replication of these two structures matched that of the trypanosome flagellum and basal bodies respectively. As the cell cycle proceeded, increased staining of the cell body was observed. Quantitation of fluorescence levels revealed that mitotic cells possess double the amount of TbRab23YFP protein compared to post-mitotic cells, consistent with flagellar replication (Figure 2C). The location of TbRab23HA mirrored TbRab23YFP (Figure 2D, right panel). We consider the possibility that two tags caused TbRab23 to mislocalize to the same structures to be highly unlikely. Furthermore, both of the vectors used generated modest overexpression, of the order of five-fold . Similar staining has been reported for the trypanosome orthologue of IFT172 , which also has a role in Shh signaling in mammals . Immunofluorescence (IF) using anti-HA and anti-GFP on untransfected cells did not produce fluorescence signal (data not shown). In addition, tetracycline-inducible RNA interference (RNAi) was used to validate the specificity of the HA-tag in BSF and PCF cells. Most significantly, no nuclear staining was observed with either tagged form of TbRab23 or at any point in the cell cycle. Given the small size of the HA epitope we consider it highly unlikely that this obscured a nuclear localization signal. Furthermore, fusion of GFP adjacent to the NLS of an authentic nuclear envelope protein, NUP-1, did not prevent correct nuclear targeting of that protein, suggesting that the presence of a tag does not necessarily disrupt nuclear import (data not shown).
To confirm the relationship of TbRab23 with the flagellum, cells were detergent extracted to remove cell membrane and soluble proteins, followed by depolymerization of the subpellicular array, leaving the flagellum cytoskeleton intact . This procedure purifies the axoneme, flagellum attachment zone (FAZ), paraflagellar rod (PFR), four specialized microtubules and the basal body from remaining cellular components.
When the equivalent TbRab23 chimeras were expressed in procyclic cells the localization was similar to the bloodstream stages, suggesting that the location of TbRab23 is conserved between developmental life forms (Figure 5A). Subsequently, the requirement for TbRab23 in parasite proliferation was examined in BSF and PCF cells by RNAi. TbRab23 was found to be non-essential in BSF cells (Figure 5B, top panel). Knockdown efficiency was validated by Western blot on a TbRab23RNAi mutant ectopically expressing TbRab23HA. Expression of TbRab23 protein was rapidly suppressed by ~95% at 24 hours whilst TbBiP levels remained unchanged indicating that the RNAi construct was specific and very efficient. Examination of cells throughout induction by IF showed the transient appearance of a small population (6%) with an abnormal number of kinetoplasts and nuclei, but these cells were rapidly lost from the population (data not shown), and no significant change to cell cycle progression was recorded thereafter. No proliferation defect was noted in PCF cells (Figure 5B, bottom panel). Knockdown efficiency was validated by Western blot on a TbRab23RNAi mutant ectopically expressing TbRab23HA revealing that recombinant protein levels were reduced by ~90% after 24 hours of induction.
Rab23 was originally descibed in metazoan cells as being a component of a signaling pathway important to development, and this model is now well supported [7–11]. However, the wide taxonomic distribution of Rab23 amongst diverse unicellular organisms indicates a more fundamental cellular role for this protein beyond regulation of multicellular development, and is supported, for example, by evidence that Rab23 is not part of Shh signaling in Drosophila melanogaster . Identifiable homologues of Patched or Smo are absent from trypanosomes and a great many other taxa, making the existence of a primitive Shh-like signaling network unlikely.
Comparative genomics reveals a strong correlation between possession of a motile flagellum and a Rab23 gene, supporting an early evolving flagellar-related function and also suggesting evolutionary pressure to retain Rab23 when the motile flagellum is present. Hence Rab23 has likely been repeatedly lost throughout evolution in organisms where the motile flagellum has been discarded, further supporting a model where the primary function(s) of Rab23 are flagellar-associated. G. lamblia and P. falciparum, are apparent exceptions, but both organisms possess divergent flagella; P. falciparum has lost IFT proteins and flagellum biogenesis is fully IFT-independent, whilst in G. lamblia, flagellum biogenesis is fully or partially IFT-independent [44, 45].
Epitope-tagged TbRab23 localizes to the flagellum and cell body in BSF and PCF cells, and subcellular fractionation and immuno-gold labeling confirmed the presence of flagellar-associated TbRab23. These data support a role for Rab23 in flagellum function in trypanosomes, consistent with the localization of mammalian Rab23 in MDCK cells and its role in ciliary biogenesis and protein turnover [10, 19, 20, 37]. Clearly these data are at variance with Dhir and Field (2004). However, the affinity-purified antibodies generated for this study do detect several high molecular weight antigens in trypanosome cell lysates in addition to TbRab23, indicating the presence of cross-reactive antibody. This is cause for concern as many nuclear components are high molecular weight coiled-coil proteins, including NUP-1 . Moreover as NUP-1 is a high abundance repetitive protein, immunostaining may have emphasized the cross-reactive component of this and other nuclear envelope antigens.
Concurrent with a potential role in flagellar function TbRab23 exhibits features that are similar to IFT proteins. Firstly, TbRab23 extends beyond the PFR into the cell and is likely present on the flagellar membrane, as TbRab23 along the length of the flagellum is extractable by detergent, similar to many trypanosome IFT proteins . Secondly, the presence of TbRab23 at the basal body is significant, as basal body transitional fibers act as docking sites for IFT particles [46, 47] and many IFT proteins additionally localize to this area . Thirdly, detergent extraction abolished most of the TbRab23 flagellar signal, whilst the basal body signal was retained. Differential extractability has previously been reported for IFT172 , IFT20  and proteins of the BBSome . Therefore, in addition to localization, TbRab23 shares biochemical properties with proteins important in ciliary maintenance and biogenesis. Finally, subcellular fractionation of Chlamydomonas reinhardtii flagella identified Rab23 in the membrane and matrix fraction, together with IFT A and B complexes . TbRab23 was not found in a trypanosome flagellum proteome , presumably because flagellar membranes were stripped before analysis. However we are unable to unequivocally assign TbRab23 to the flagellum membrane or other flagellar subcompartment from the data presented here.
Surprisingly, depletion of TbRab23 did not result in a flagellar defect. This is unlikely due to inefficient suppression of protein levels, as the BSF and PCF RNAi construct effectively eliminated recombinant TbRab23 expression at levels five to ten-fold higher than endogenous protein . Rather Rab23 may not be required for construction or maintenance of the flagellum in trypanosomes. Alternatively TbRab23 may exhibit functional redundancy with other IFT proteins or participate in a role that is not essential for the continuation of the cell cycle.
The data reported here indicate that trypanosome Rab23 is associated with the flagellum, and not with the inner nuclear envelope as previously suggested. Use of two distinct tags, correlation between possession of a Rab23 and a flagellum, cross-reactivity in the original sera, plus additional evidence from mammalian cell studies argues that this is the correct assignment. Hence these data unify the location of Rab23 across the eukaryotes.
flagellum attachment zone
Ra s-related proteins in b rain, Shh: Sonic hedgehog.
We are grateful to the following for generous gifts of reagents: James Bangs (antibodies against TbBiP), Keith Gull (antibodies to TbPFR and FAZ) and Joel Dacks (Rab5 sequences). We also thank the members of our laboratory for many discussions and support. This work was funded by the MRC (studentship to JHL) and the Wellcome Trust (project and program grants to MCF).
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.