|Home | About | Journals | Submit | Contact Us | Français|
FKBP12 proteins are able to inhibit TOR kinases or calcineurin phosphatases upon binding of rapamycin or FK506 drugs, respectively. The Trypanosoma brucei FKBP12 homologue (TbFKBP12) was found to be a cytoskeleton-associated protein with specific localization in the flagellar pocket area of the bloodstream form. In the insect procyclic form, RNA interference-mediated knockdown of TbFKBP12 affected motility. In bloodstream cells, depletion of TbFKBP12 affected cytokinesis and cytoskeleton architecture. These last effects were associated with the presence of internal translucent cavities limited by an inside-out configuration of the normal cell surface, with a luminal variant surface glycoprotein coat lined up by microtubules. These cavities, which recreated the streamlined shape of the normal trypanosome cytoskeleton, might represent unsuccessful attempts for cell abscission. We propose that TbFKBP12 differentially affects stage-specific processes through association with the cytoskeleton.
African trypanosomes are extracellular protozoan flagellated parasites responsible for sleeping sickness in humans and nagana in cattle. The life cycle of Trypanosoma brucei encompasses different stages, including the long slender bloodstream forms (BF) proliferating in mammalian blood and the procyclic forms (PF) that actively multiply in the gut of the Glossina vector (1).
Trypanosomes are among the most divergent eukaryotes in evolution and display specific features, many of which are related to cell division probably due to the fact that most organelles are present at one copy per cell and have to be duplicated and segregated synchronously between the daughter cells. This division involves check points that differ from those of other eukaryotes, such as the control of karyokinesis when cytokinesis is inhibited (2, 3) and vice versa (4). Molecular effectors of these check points, such as mitogen-activated protein kinase and cyclin-dependent kinase, are present in trypanosomes but diverge in function compared to other eukaryotes (5, 6).
The flagellum and its motility appear to play a key role in the control of cell division (7–9). This organelle initiates at the basal body, which is associated to the kinetoplast (10, 11), emerges from the flagellar pocket (FP), and it is attached along the cell body for most of its length by the flagellum attachment zone (FAZ). The flagellum contains a canonical axoneme and the paraflagellar rod (PFR) that are physically linked (12–14). The duplication and segregation of these structures are interdependent. During cytokinesis, the ingression of the cleavage furrow follows an axis in between the new and the old flagellum. The position and initiation of the furrow are closely related to the FAZ, as demonstrated by the study of flagellum mutants (15–21).
In eukaryotes such as yeasts or mammals the TOR pathway is a major player in the control of cell division mediated by the action of two protein complexes, TORC1 and TORC2 (22–25). These complexes contain the two different threonine/serine kinases TOR1 and TOR2 in the yeast Saccharomyces cerevisiae (26–28), and one TOR protein in mammals (29). TORC1 complex controls cell mass (25, 30–32) and TORC2 the spatial aspects of cell division through cytoskeleton formation (33, 34). The role of the TOR pathway was uncovered through its inhibition by rapamycin (35). This drug, as well as a compound termed FK506, binds a cytoplasmic protein termed FKBP12 (for FK506 binding protein of 12 kDa). Binding of these compounds to FKBP12 suppresses the enzymatic peptidylprolyl cis/trans isomerase (PPIase) activity of the protein (36, 37). The rapamycin/FKBP and FK506/FKBP then form ternary complexes with TOR and calcineurin, respectively (29, 30, 38, 39), leading to the inhibition of the downstream signal transduction pathways. FKBP12 binds and modulates the activity of several intracellular targets, such as the calcium channels ryanodine receptor (40) and inositol 1,4,5-triphosphate receptor (41, 42).
In trypanosomes, two TOR proteins have been identified (43–45). In BF, their respective functions seem to match those found in other eukaryotes. They are part of two different protein complexes with different cellular localizations. Gene knockdown of TbTOR1 resulted in reduced cell growth and arrest in G1 concomitant with reduced protein synthesis, whereas TbTOR2 RNA interference (RNAi) induced abnormal morphology and cytokinesis defects generating cells with multiple flagella and nuclei. Finally, rapamycin inhibited cell growth through interference with TOR2 but not TOR1 formation. Recently, two novel TOR kinases, TbTOR3 and TbTOR4 (formerly TbTOR-like 1 and TbTOR-like 2) were identified in the genome of T. brucei (43). TbTOR3 is a cytoplasmic TOR kinase involved in polyphosphate metabolism, acidocalcisome maintenance (46), and virulence (47). TbTOR4 is involved in differentiation of slender into stumpy bloodstream forms (48).
We identified four members of the FKBP family and investigated their role in T. brucei. The knockdown of FKBP12 was the only one to trigger detectable effects on trypanosomes in culture, with reduced motility in PF and impaired cytokinesis in BF. Remarkably, in BF the depletion of TbFKBP12 was associated with the presence of internal inside-out translucent cavities seemingly resulting from unsuccessful attempts of cell division. We propose that TbFKBP12 is a cytoskeleton-associated protein with BF-specific function in the FP region.
The double-marker PF cell line 29-13 (49) was cultured at 27°C in SDM 79 medium containing 10% fetal calf serum. Cell density was kept at 1 to 10 million cells/ml by regular dilution, and growth curves are presented as cumulative data. Parasites were transfected as described previously (50). Briefly, cells were transfected with 5 μg of the linearized plasmid using a Bio-Rad Gene Pulser II. Cell lines were obtained by clonal dilution under selection with 5 μg of phleomycin/ml. For the induction of RNAi, cells were grown in the presence of 1 μg of fresh doxycycline (DOX)/ml.
The single-marker BF cell line 328-114 (49) was cultured at 37°C and 5% CO2 in HMI-9 medium containing 10% fetal calf serum without tetracycline (Tet system approved fetal bovine serum; Clontech) (51). Cell density was maintained at 0.1 to 1 million cells/ml by regular dilution, and growth curves were presented as cumulative data. Parasites were transfected as described previously (52). Briefly, cells where transfected with 10 μg of the linearized plasmid using the AMAXA human T-cell nucleofector. Stably transfected clonal cell lines were recovered by selection with 1 μg of hygromycin/ml (RNAi) or with 1 μg of phleomycin/ml (complementation) in a 24-well plate. For induction of RNAi, cells were grown in the presence of 1 μg of fresh DOX/ml.
To obtain anti-FKBP12 antibodies, recombinant GST-TbFKBP12 was cloned in the pGEX4T1 vector (Amersham) and overexpressed in bacteria, purified in phosphate-buffered saline (PBS) on glutathione-Sepharose 4B (Amersham) beads and then injected into a mouse to produce polyclonal antisera.
The following primers were designed using Beacon Designer 7.0 (Premier Biosoft, CA) to PCR amplify a 331-nucleotide fragment of the 339 nucleotides, the FKBP12 gene from wild-type T. brucei genomic DNA: TbFKBP12 F (5′-ATATGGATCCAAATGACTGCGTTATGATGG-3′) and TbFKBP12 R (5′-ATGCCTCGAGTCAAACAACATCCAACAGCG-3′). PCR product was cloned in the p2T7TiTA177 vector (53). The sequence of the insert was verified by DNA sequencing (Genome Express), and the plasmid was then linearized by digestion with NotI before transfection.
The chemical composition of a recoded open reading frame (ORF) coding for the TbFKBP12 protein was obtained from Entelechon. This sequence is synonymous to the T. brucei TbFKBP12 gene but highly mutated: it has the same amino acid sequence but a different nucleotide sequence and is therefore not targeted by the RNAi machinery of the cell. The sequence (CTCGAGATGAGTAGGAACGATTGTGTGATGATGGATAAAATAATCGAAGGTGACGGTAAGACAATTCCTAGGCAAGGATCAATTGTTACGTTAGATTACGTGGGGTACTTACCAGACGGAAGGAAATTTGATAGTACCATCGAGCGTGGAAAACCATTCGTGTTTAGGGTAGGTTGTGGCGAAGTGATTAAGGGTTGGGATGAAGGGATCGTTCAAATGAGCAAAGGAGAAAGGTCGAGATTGACTATGCCCCCGAGCTTAGCATTCGGTAGTACAGGATTTCCAGGGATTATCCCACCGAATACGGTGATAGTTTTCGAAGTTACATTACTCGACGTGGTCTAGAAGCTT) was cloned in a pTSA-rib vector (54) using the primers TbFKBP12-rev F (5′-ATATCTCGAGATGTCGCGAAATGACTGCG-3′) and TbFKBP12-rev R (5′-ATATAAGCTTTCAAACAACATCCAACAGCG-3′). This plasmid is targeted to a tubulin locus and allows a constitutive expression under the control of the ribosomal DNA promoter. The plasmid was linearized with BglII before transfection into the TbFKBP12RNAi cell line.
After RNAi induction RNAs from the BF and PF cells after RNAi induction were extracted with TRIzol (Invitrogen) and used as a template for reverse transcription with RT Superscript II (Invitrogen). The cDNAs were then amplified by quantitative reverse transcription-PCR (qRT-PCR) with specific primers for the TbFKBP12 and histone H2B mRNAs using the SYBR green mix (ABgene).
For TEM, samples were fixed overnight at 4°C in 2.5% glutaraldehyde–0.1 M cacodylate buffer (pH 7.2) and postfixed in OsO4 (2%) in the same buffer. After serial dehydration samples were embedded in Agar 100 (Agar Scientific, Ltd., United Kingdom) and left to polymerize at 60°C. Ultrathin sections (50 to 70 nm thick) were collected in Formvar-carbon-coated copper grids using a Leica EM UC6 ultramicrotome and stained with uranyl acetate and lead citrate. For immunogold detection by ultrathin cryosectioning, cells were fixed in 3% paraformaldehyde–0.5% glutaraldehyde–0.1 M cacodylate buffer (pH 7.2), embedded in 10% gelatin and 2.3 M sucrose, and frozen in liquid nitrogen. Sectioning of frozen samples was done on a Leica EM UC7 ultramicrotome. Sections on carbon-Formvar grids were probed with a mouse anti-TbFKBP12 antibody and an anti-mouse gold conjugate (5 nm) and then mounted in methylcellulose–1% uranyl acetate films. Observations were made on a Tecnai 10 electron microscope (FEI) and images were captured with a MegaView II camera and processed with AnalySIS and Adobe Photoshop software. For scanning electron microscopy (SEM), samples were fixed overnight at 4°C in 2.5% glutaraldehyde–0.1 M cacodylate buffer (pH 7.2) and postfixed in 2% OsO4 in the same buffer. After serial dehydration, the samples were dried at a critical point and coated with platinum according to standard procedures. Observations were made in a Tecnai FEG ESEM Quanta 200 (FEI). For detergent extraction, the cells were treated with 1% Triton X-100 at 4°C in PBS for 10 min to strip the plasma membrane and to visualize the parasite cytoskeleton. Samples were washed twice in PBS, fixed in glutaraldehyde, and processed for SEM under standard conditions (55).
Trypanosomes lysates, detergent-soluble proteins, and cell cytoskeletons were prepared as described previously (56) and Western blotted as described previously (57). The primary antibody dilutions were as follows: monoclonal anti-trypanin, 1/2,500 (58); anti-enolase, 1/50,000 (59); and mouse polyclonal anti-TbFKBP12, 1/50.
Cells were spread on poly-l-lysine-coated slides and either fixed and permeabilized in methanol at −20°C or fixed in 3.7% formaldehyde and then permeabilized in 0.01% Triton TX-100 before processing as described previously (60). Indirect immunofluorescence analyses were performed with mouse anti-TbFKBP12, anti-PFR2 antibody L8C4 (61), anti-FAZ antibody L3B2 (61), anti-acetylated α-tubulin (MAb 6-11b-1; Sigma), anti-TBBC antibody (basal body) (62), anti-enolase antibody (59), anti-trypanin antibody (58), and anti-BILBO1 antibody (63). The FP was labeled with Texas Red tomato-lectin (Vector Laboratories) (64). DNA was stained with DAPI (4′,6′-diamidino-2-phenylindole). Images were captured with a Zeiss Axioskop 2 coupled to a charge-coupled device camera (AxioCam HRM; Zeiss).
For each stage, motility analysis was performed from two separate experiments. For each time point (0, 24, 48, and 72 h of induction), 16 to 20 movies were recorded (200 frames, 50 ms of exposure). Samples were observed in their respective media (0.5 × 106 BF/ml of HMI-9 or 5 × 106 PF/ml of SDM 79) under the ×10 objective lens of an inverted DMI4000 Leica microscope coupled to a Retiga-SRV camera (QImaging). Movies were converted with the mpeg Streamclip V.1.9b3 software (Squared 5) and analyzed using MedeaLAB CASA tracking V.5.5 software (Medea AV). For each movie, up to 199 cells were simultaneously tracked in silico. The studied parameters were as follows: curvilinear velocity (VCL; speed along the real path) and straight line velocity (VSL; speed calculated along the virtual linear path from the starting to the ending point). The data were transferred and compiled in Microsoft Excel, and statistical analyses were performed using KaleidaGraph v4.0 software (Synergy Software). The data were first filtered (number of frames, >197; 1% < linearity < 99%; VSL > 0.1) to exclude incomplete and/or false. An analysis of variance test was then performed to study the effect of the RNAi induction over time on all of the measured parameters (Tukey ad hoc post test at 0.5 for intergroup comparisons; P < 0.001).
For a more precise analysis, ~20 movies of individual cells (200 frames, 100 ms of exposure) were recorded at different times after induction (0 and 48 h for BF; 0 and 72 h for PF). Samples were observed and processed as described above under a 100× NA 1.4 Plan Apo objective lens. A qualitative observation of each time-lapse movie was then performed to characterize flagellar beatings (wave orientation, beat amplitude, beat efficiency, and coordination of flagella).
In T. brucei, computer-assisted BLAST analysis identified four matching sequences for the FKBP12 protein (see Fig. S1 in the supplemental material). Because of its size and highest level of identity, we focused in the present study on Tb927.7.3420, the putative homologue of FKBP12, which we termed TbFKBP12. At the amino acid level, TbFKBP12 shares 53 and 51% identities with the S. cerevisiae and the human homologue, respectively. The predicted 12.2-kDa protein (112 amino acids) contains a typical binding domain for the FK506 drug (FKBD), as well as a potential catalytic site for peptidylprolyl cis-trans isomerase (PPIase) activity (see Fig. S2 in the supplemental material). Despite the fact that the large majority of FKBPs display in vitro PPIase activity, we never succeeded to detect this activity in recombinant TbFKBP12 using classical synthetic prolyl-oligopeptide substrates (data not shown). The lack of detectable enzymatic activity could be related to the presence of several substitutions on key catalytic residues in the drug-binding pocket of TbFKBP12 that were proposed to mediate PPIase activity in other FKBPs (65) (asterisks in Fig. S2 in the supplemental material). Nevertheless, nuclear magnetic resonance (NMR) studies showed that the recombinant protein was correctly folded under the assay conditions and able to bind FK506 and rapamycin but apparently not the proline substrates, even to a concentration of 100 μM close to the Km for these peptides (290 μM; R. Aido-Machado, D. Salmon, and J. R. Pires, unpublished data). In addition, surface plasmon resonance analysis of the recombinant TbFKBP12-rapamycin interaction revealed an apparent KD (equilibrium dissociation constant) of 6.8 nM, slightly higher than that determined for the human FKBP12 (1.6 nM) (66; K. M. S. Kabral and D. Salmon, unpublished data). Although the overall structure of the FKBP is conserved, we speculated that the substitutions of some residues in TbFKBP12 including (Y87F and H92F) might be important in hydrogen bonding interaction within the active site of the enzyme (65), the lack of hydrogen bonds in both Phe substitutions could make the binding weaker with the peptide resulting in the loss of detectable PPIase activity (Aido-Machado et al., unpublished).
Antibodies were produced against recombinant TbFKBP12 fused to a glutathione S-transferase (GST; N-terminal) tag. These antibodies specifically recognized a 12-kDa protein in T. brucei total cell extracts (lanes L, Fig. 1A). After cell fractionation, TbFKBP12 was found to be mainly associated with the cytoskeleton (lanes CK, Fig. 1A), while the control cytoplasmic enzyme enolase (59) and flagellar protein trypanin (58) were found in their respective expected fractions (Fig. 1A). Further experiments conducted to determine whether TbFKBP12 was associated with the cytoskeleton via subpellicular microtubules yielded inconsistent results (data not shown). Using indirect immunofluorescence assays (IFA), we found TbFKBP12 in a compartment juxtaposing the FP in BF (Fig. 1B). Frequently, the signal corresponding to TbFKBP12 in 1K1N cells had a bilobular aspect centered along the flagellum axis (Fig. 1B, arrow). Further colocalization experiments using tomato lectin as a marker for the FP lumen (64) and antibodies directed against BILBO1, an FP collar protein (63), confirmed the FP localization (Fig. 1C and andD).D). Furthermore, these markers allowed us to define the relative position of TbFKBP12 at the posterior end, adjacent to BILBO1 and toward the interior of the cell. Using immunogold transmission electron microscopy (TEM) analysis on ultrathin cryosections, the gold label was often associated to an unidentified membrane network close to the posterior end of the cell and near the FP (see Fig. S3 in the supplemental material). In these cryosection pictures, the identification of the FP region was based on the presence of the axoneme, the PFR, the flagellum, FP membranes, and the FP collar. In contrast to membranes, cytoskeleton structures are poorly defined on cryosections. Despite numerous efforts, no consistent signal was detected in PF either by IFA or by immunogold analysis (data not shown).
RNAi analysis was performed on PF and BF expressing the T7 phage RNA polymerase and the bacterial TET repressor (53). Knockdown of TbFKBP12 was accomplished by the tetracycline-dependent expression of a TbFKBP12-specific double-stranded RNA (dsRNA) encoded on the p2T7TiTA177 vector. Among the four T. brucei FKBP-related genes, only TbFKBP12 knockdown led to a detectable phenotype. After RNAi-mediated reduction of both mRNA and protein TbFKBP12 levels (Fig. 2A to toD),D), the growth of PF stopped after 3 days and that of BF after 1 day (Fig. 2). TbFKBP12 seemed to be less stable in PF than in BF, since after 3 days of RNAi induction the mRNA levels decreased to 81% in PF and to 78% BF, but the protein was still present in BF while not detected any longer in PF (Fig. 2A to toD).D). To verify the specificity of the RNAi growth phenotype, we did a complementation analysis using a recoded sequence of TbFKBP12 ORF called FKBP12-rev, which should be insensitive to RNAi but preserves the wild-type (WT) ORF. RT-PCR demonstrated efficient depletion of the endogenous TbFKBP12, whereas TbFKBP12-rev was strongly expressed and insensitive to RNAi as expected (see Fig. S4A in the supplemental material). This expression restored the normal growth (see Fig. S4B in the supplemental material), confirming the specificity of the phenotype.
To further define the growth-arrest phenotype upon knockdown of TbFKBP12, we counted the number of nuclei and kinetoplasts per cell after DAPI staining. The cell cycle progression in trypanosomes is easily followed from G1 phase with 1 kinetoplast (K) and 1 nucleus (N), to G2 phase with a 2K1N configuration that progresses into mitotic cells with 2K2N, which then splits into two daughter cells after cytokinesis (67).
In PF, there were no major changes in the relative proportions of 1K1N, 2K1N, or 2K2N normal populations after 24 h of RNAi (Fig. 3A). However, all of these populations decreased progressively between 24 and 120 h, concomitantly with the emergence of cells with more than 2K2N whose abundance represented 20% of the population after 120 h of depletion.
In BF, the percentage of 1K1N and 2K1N cells remained relatively constant between 12 and 36 h after DOX addition, whereas the 2K2N population doubled during the first 6 h and then decreased to come back to its original value (Fig. 3B, light gray). During the same period we observed the emergence of a population of abnormal cells with more than 2K2N, which represented half of the population after 48 h of RNAi induction. The accumulation of more than 2 K and 2 N per cell indicated genome duplication but failure to complete cytokinesis. Thus, the growth arrest in BF was due to a failure in cytokinesis, in contrast to the PF situation. With a doubling time of 15 h for PF and 8 h for BF, six cell cycles correspond roughly to 90 h for PF and 48 h for BF. At these times, 21% of the PF population and 55% of the BF population were monster cells.
RNAi silencing of TbFKBP12 specifically led to a motility defect in PF. This was first investigated by a standard sedimentation analysis (18). Upon depletion of TbFKBP12, PF sedimented at the bottom of the test flask within 4 h, which did not occur in BF (Fig. 4A and andB).B). This motility defect was not responsible for the growth phenotype of PF, since shaking the culture did not reverse it, in contrast to other RNAi phenotypes causing reduced motility (8, 9; data not shown). Video microscopy revealed that the flagella of these cells were still beating, but this movement was not dragging cells in any particular direction (see Movie S1 to S3 in the supplemental material). In trypanosomes, the propulsive wave initiated from the anterior (tip) to the posterior (base) end of the flagellum drags the parasite forward and propagates with high frequency and low amplitude, whereas the reverse, base-to-tip wave, is used for changes in direction and propagates with low frequency and high amplitude (8, 9, 68). We analyzed several parameters for both PF and BF: (i) the curvilinear velocity (VCL, or speed along the real path), (ii) the straight-line velocity (VSL, or speed calculated along the virtual linear path from the starting to the ending point), (iii) the trajectory linearity (percentage of the total path), and (iv) the frequency of long and short stops. In PF all of these parameters progressively decreased along the time of TbFKBP12 depletion (Fig. 4C and data not shown). After 72 h, even though beats in both directions were observed, reverse waves (base to tip) were predominant, and propagation of the beats was often incomplete. In BF, no significant modifications were observed on these parameters up until 48 h after triggering RNAi (see Movies S4 to S7 in the supplemental material), when the velocity and distance covered on the whole decreased, although the linearity of the motility remained unchanged (Fig. 4D and data not shown). These effects were correlated with the increase in the number of multiflagellated cells with uncoordinated beating of their flagella and are in agreement with the sedimentation data. Overall, these results indicated that in PF the propulsive wave was affected, imposing a handicap on the parasite motility, whereas in BF reduced motility seemed mainly due to uncoordinated beating of multiple flagella in monster cells.
Since cytokinesis is affected by defects in flagellum biogenesis (15–19), we analyzed the flagellum in TbFKBP12-depleted BF and PF cells. In both cell types, the formation of the paraflagellar rod (marked by an anti-PFR2 antibody), basal bodies (marked by the anti-TBBC antibody), and flagellar attachment zone (FAZ, marked by the anti-FAZ1 antibody) appeared normal upon depletion of TbFKBP12 (Fig. 5A). These images also showed the high number of flagella present in BF after 48 h of RNAi induction, paralleled by the accumulation of kinetoplasts and nuclei. The dual presence of flagellum-BB was still preserved, as well as that of the flagellum and the FAZ (Fig. 5A and data not shown).
In order to explore further a possible cause for the effect on motility, we used TEM and SEM for morphological analysis on FKBP12RNAi cells. TEM ultrastructural analysis confirmed the integrity of the flagella: the axoneme, the PFR, and the FAZ components were still present in FKBP12RNAi cells and showed the same inter-relationship as in WT cells (Fig. 5B). The 9+2 structure of the axoneme was preserved, together with the dynein arms and radial spokes.
We next analyzed the flagellum attachment to the cell body, since partial or complete detachment of the flagellum is one of the reported causes for lack of productive motility (69, 70). Using SEM and TEM, we did not observe alteration of flagellar attachment or anchorage to the cell body in FKBP12RNAi cells (data not shown). SEM cytoskeleton analysis of these parasites after membrane removal by detergent extraction (55) demonstrated several things. (i) The posterior region of the flagella appeared to be deeply encrusted in the cell body. (ii) Multiple cells were observed with unsegregated subpellicular microtubule (SPMT) networks, which were often associated at the middle of undivided cells (see Fig. S5C, panels 2 and 3, in the supplemental material). (iii) Some cells appeared to have abnormally separated from the posterior end or from both the anterior and the posterior ends. (iv) Finally, a low frequency of multiflagellated cells was observed with a complex and apparently unresolved SPMT network (see Fig. S5C, panel 3, in the supplemental material). Overall, these findings suggest that problems in flagellum attachment or positioning cannot account for the effects on motility of FKBP12RNAi PF cells and that cytoskeleton architecture is affected in these cells.
Video microscopy on live BF arrested in cytokinesis showed that monster cells had a rosette-like morphology bearing several cell bodies of various sizes connected together at the posterior end (Fig. 6A). For simplicity, we refer to these monster cells as “aggregates,” even though this does not mean individual cells that stuck together (see below). Each component of these monster cells still retained the streamline shape of normal cells and had one or several nonsynchronized motile flagella, and their body size was proportional to the number of flagella, nuclei, and kinetoplasts (Fig. 6A). By SEM we observed a rise in the number, size, and complexity of cell aggregates along the RNAi induction time (Fig. 6). The individual cells in these aggregates had a streamlined normal aspect, often with more than one flagellum (Fig. 6B, panels 2 to 4). The fact that the number and complexity of cell aggregates increased along the time of TbFKBP12 RNAi induction indicated that these correspond to the rosette configurations observed by light microscopy. In TbFKBP12-depleted BF the flagellum exited at odd places along the cell body (Fig. 6B, panel 3, arrow). In contrast to PF, the flagellum was rarely detached from TbFKBP12-depleted BF. TEM analysis of cross-sections of TbFKBP12RNAi cells confirmed that the cell aggregates observed by light microscopy and SEM corresponded to a single massive undivided cell (Fig. 6C). In every instance one could discern the cell components in each of the multicellular bodies, since each one had the streamlined shape characteristic of trypanosomes and a peripheral flagellum transversal section (Fig. 6C, arrows). Altogether, these observations suggested that unresolved cell aggregates corresponded to late cytokinesis-arrested cells.
Intriguingly, TEM analysis of ultrathin sections revealed the presence of “cavities” at the interior of cytokinesis-arrested cells. These cavities had a translucent lumen limited by a wall of electron-dense material, of the same thickness and characteristics as the variant surface glycoprotein (VSG) coat, and an underlying array of microtubules equivalent to the SPMT (Fig. 7A, panel 1). These cavities of variable diameters (ranging from 27 to 986 nm, see Fig. S6A and B in the supplemental material) could be readily observed after 36 h (35%, n = 82) and 48 h (40%, n = 225) of RNAi induction in cross- and longitudinal sections. The cross-sections of the cavities were circular since their horizontal and vertical diameters were highly similar (see Fig. S6A in the supplemental material). In longitudinal sections, the length of cavities ranged from 87 to 3,834 nm (see Fig. S6C in the supplemental material). At 36 h of RNAi induction the cavities seemed to be longer and larger than at 48 h (see Fig. S6D in the supplemental material). The diameters of longitudinal sections were bigger than those measured on cross-sections (see Fig. S6A and B in the supplemental material). Statistical analysis of the size versus frequency distribution indicated that this was not Gaussian. Frequently, we observed several cavities of various diameters in the same cross-section, clearly following a common axis perpendicular to the central axis of the individual cavities (Fig. 7A, panel 2). Some cross-sections between two undivided cells showed strange cytokinesis configurations (Fig. 7B, panels 1 and 2, pointed lines). In several instances, the diameter of the cavities increased toward the cell interior; cavities of the smallest size were found near the cell surface (Fig. 7C, panels 1 to 4). In the example shown, we could clearly identify the VSG coat surrounded by 20-nm cross-sectioned microtubules (MT) (Fig. 7C, panels 1 to 3, red arrows), associated with membranes resembling those of the smooth endoplasmic reticulum (SER) surrounding the MT quartet of the FAZ (Fig. 7C, panel 4, white arrow). In some cases, there was a direct link between the ER from the FAZ and the MT of the cavities (Fig. 7C, panels 4 and 5, asterisks). This MT/ER association was also observed in longitudinal sections (Fig. 7D), and it was particularly abundant at the posterior end of the cell (see below).
Since we had observed MT/ER associations particularly abundant at the posterior end of the RNAi BF cells (16.1% of cells with internal MT after 48 h of RNAi induction; n = 118) (Fig. 8A), we probed these cells for the presence of acetylated α-tubulin, which has been associated to MT biogenesis in BF (71). By IFA, we observed an increase (~10-fold, based on fluorescence intensity) of acetylated α-tubulin at the posterior end of FKBP12RNAi cells after 12 h of RNAi induction (70% of cells; n = 108) (Fig. 8B).
Despite obvious phenotypic consequences of TbFKBP12 knockdown in both BF and PF, this protein did not show PPIase activity by standard assays. Since immunoprecipitation of the endogenous protein did not provide enough material, we could only use the recombinant version of TbFKBP12 to monitor its activity. This restriction cannot simply account for the lack of activity because NMR studies showed that the recombinant protein used for the assays is correctly folded and binds rapamycin and FK506 similarly to a control human protein. Therefore, PPIase activity might not be required for the physiological function of TbFKBP12, as was shown in both yeast and a pathogenic bacterium, Burkholderia pseudomallei (72, 73).
Proteins associated with FKBP12 have already been characterized in T. brucei, namely, TbTOR1, TbTOR2, TbTOR3, and TbTOR4 (43–48). TOR, FKBP, and rapamycin are in a tripartite complex, therefore interfering with TOR or FKBP could lead to similar phenotypes. Although TbTOR1 and TbTOR3 were found to be involved in cell cycle regulation and acidocalcisome maintenance, respectively, TbTOR2 was proposed to control cytokinesis and TbTOR4 was proposed to control slender to stumpy differentiation. However, the TbTOR2 knockdown phenotype differed from that exhibited by TbFKBP12RNAi cells. We did not observe the FP swelling phenotype, while the intracellular cavities of the TbFKBP12RNAi cells were not observed in previous studies (45).
Fractionation experiments on detergent-extracted cells demonstrated the presence of TbFKB12 in the cytoskeleton fraction. However, we could not identify the cytoskeleton component to which TbFKBP12 is associated. In BF, TbFKBP12 was localized to a region adjacent to the FP collar, where it might be associated with an unidentified membrane compartment. Several cytoskeleton components are located near the FP zone, namely, the basal bodies and associated filaments, the MT quartet associated to the FAZ and the FP collar (74). The fenestrated endoplasmic reticulum (FER) identified by TEM tomography (74, 75) is a good candidate for the TbFKBP12-containing compartment, but this hypothesis is difficult to assess in the absence of specific marker to label the FER. We propose that the bulk of TbFKBP12 is associated with the cytoskeleton in both PF and BF, whereas only in BF would a fraction be located in the FP region. The higher local concentration of TbFKBP12 at the FP region of BF would account for its detection in this location only, whereas its absence from the FP of PF would explain the differential RNAi phenotype between PF and BF.
While FKBP12 is nonessential in yeast (76), TbFKBP12 was required for both PF and BF growth. In neither PF nor BF was biogenesis of the flagellum or of any other cell organelle analyzed affected. Regarding the flagellum, all of its main components, as analyzed by IFA and EM, appeared to be normal in FKBP12RNAi cells. In PF the loss of TbFKBP12 affected motility, with a quick drop in curvilinear velocity due to a loss of coordination in the production of propulsive and reverse waves. This is translated in the rapid cessation of productive movement that causes cells to sediment in the culture flask within 60 h of RNAi induction. We ignored the primary cause for this effect on motility, but it does not seem to be related to an obvious loss of flagellar components or to flagellar detachment or flagellum anchoring. Calcium movement is a process that could be targeted very fast. Within this context, flagellar tip-to-base and base-to-tip waves have been shown to depend on different calcium concentrations on isolated flagella from the trypanosomatid Crithidia oncopelti (77). Furthermore, Ca2+ movement in sperm flagella and cilia is controlled through the modulation of IP3R and RyR in the ER, both ligands of FKBP12 (78, 79). Another possibility is that cytoskeleton organization (flagellum and/or SPMT) in TbFKBP12RNAi PF might indeed be affected. Differences in cell body shape are correlated with a diverse range of cell behaviors contributing to the directional motion of the cell. For instance, straighter cells swim more directionally, whereas cells that exhibit little net displacement appear to be more bent (80). Because the flagellum is attached laterally along the length of the cell body, a helical wave of the flagellum necessarily applies a torsional stress to the cell body. Depending on the degree of viscoelasticity of the cell body, the cell body may well be able to store a significant amount of torsional elastic energy in reversible deformation of the microtubule cytoskeleton. This elastic energy could then be released when the flagellum started to rotate in the opposite direction (68). A disorganization of the cytoskeletal microtubule network could prevent proper storage or release of this elastic energy and consequently impair the parasite motion.
Contrary to the effects seen in PF, in BF the motility defects were only observed after long periods of RNAi induction, and they were associated with uncoordinated beating of the multiple flagella of undivided cells. Consequently, the main phenotype of BF was the inhibition of cytokinesis with no effect on mitosis, leading to the accumulation of monster cells bearing multiple flagella, kinetoplasts, and nuclei similar to previously described phenotypes linked to the invalidation of some flagellar components (8, 9, 58, 81). Given the frequency and the diversity of cell processes were this phenotype has been observed, it might be directly or indirectly related to cytokinesis (6, 71). To our knowledge, the presence of cavities limited by a wall composed of MT, a plasma membrane, and a VSG coat has not been reported in any of the cytokinesis-arrested phenotypes. Neither have we ever seen similar features in our vast collection of RNAi mutants and trypanosomes under deleterious physiological conditions, even after long injury periods (4 to 6 days). We therefore concluded that the presence of cavities is specific to the ablation of TbFKBP12 and that the effect on cytokinesis is potentially due to the disruption of cytoskeleton dynamics. The analysis of hundreds of images led us to the conclusion that the cavities are ellipsoids in shape and run along the longitudinal axis of the cell. When multiple, the cavities were not randomly distributed in the cytoplasm but rather aligned along a foresighted division axis. Cross-sections of the smallest cavities showed that their microtubules were associated to a membrane similar to the smooth endoplasmic reticulum (SER), which could provide the membrane lining the cavities. Based on all of these data, we favor a working model for the generation of these cavities (Fig. 9). VSG coats, SPMTs, and SER membranes are nucleated and formed on the cytoplasmic side of the cell surface and, from then on, they move inward at the same time as more MTs are joining to increase the cavity diameter and length. This process then progresses on the longitudinal axis of the cell to form streamlined cavities. Interestingly, SPMTs seemed to seed the formation of cavities and to leave behind a region deprived of them at the cell surface. The existence of extensive amounts of MTs bound to SER inside these cells led us to conclude that there is a very active process to generate them and that these SER-associated MTs provide the necessary MTs and membranes to generate the cavities. Accordingly, the cavities may inherit the SPMT spatial patterning from their harboring cells through the lateral seeding of newly formed microtubules on preexisting SPMTs. A similar process has been previously suggested (82). We propose that these cavities result from unsuccessful attempts of cell abscission; when multiple, they reflect an uncontrolled and redundant command for a cytokinesis event involving the construction of a surface coat (Fig. 9C). Since we consistently observed these cavities between two flagella, it is possible that two neighboring, mislocalized flagella may signal the formation of a cleavage furrow between them and promote the formation of internal MT cytoskeletons (Fig. 9B and andD).D). It has been proposed that the correct positioning of the new and old flagella is necessary to define cellular pattern (83, 84). This could occur through the flagellar connector (FC), a small structure that maintains the distal tip of the new flagellum in permanent contact with the side of the old flagellum. Within this context it is important to mention that in several instances, we observed a direct association between the FAZ ER and some of the MTs lining the cavities, which might be relevant giving the importance of the FAZ in defining cell polarity (85, 86).
How do the observed effects relate to the location of TbFKBP12 in BF? One possibility is that this protein regulates the formation of the SPMT cytoskeletons at the posterior end of BF cells. As mentioned above, TbFKBP12 might be located in the fenestrated endoplasmic reticulum, which, incidentally, resembles the membranes associated with the MTs fed to the cavities. The hypothesis that TbFKBP12 might regulate de novo SPMT cytoskeleton formation at the posterior end is indirectly supported by two observations. First, there is an accumulation of acetylated α-tubulin and microtubules at the posterior ends of RNAi BF cells. To our knowledge, the accumulation of free MTs in trypanosomes has not been reported in any situation. Second, a similar deposition of α-tubulin at the posterior end has been observed in PF for tyrosinated α-tubulin (60, 85) but not for its acetylated form (87). Importantly, the assembly of newly synthesized SPMTs occurs at the posterior ends of PF cells (60, 85, 87) and, more recently, the acetylation of α-tubulin has been associated with MT biogenesis in BF cells (71). Therefore, the acetylated α-tubulin accumulation at the posterior ends of TbFKBP12RNAi BF cells represents an early event related to MT neogenesis prior to cavity formation.
Further work is obviously needed to establish the nature of the TbFKBP12-containing region and the function of the protein and to define the architecture of the observed cavities in three dimensions by TEM tomography, which may provide clues as to how cytoskeleton formation and cytokinesis occurs in trypanosomes.
This study was supported by the Belgian Fund for Scientific Research and the Interuniversity Attraction Poles Program–Belgian Science Policy. The CMMI is supported by the European Regional Development Fund and the Walloon Region. Work at the Pasteur Institute was funded by the Pasteur Institute, the CNRS, and a grant from the ANR (08-MIE-027).
We thank Keith Gull (Oxford University, United Kingdom) for antibodies and Frédéric Fontaine, Noémie Lepage, Laurence Lecordier, and Daniel Monteyne for assistance. A.B. was financed by FRIA, and B.R. is funded by a Roux postdoctoral fellowship.
Published ahead of print 26 October 2012
Supplemental material for this article may be found at http://dx.doi.org/10.1128/EC.00077-12.